WW, WZ, and ZZ Production
Masahiro Morii, Harvard University
SM@LHC 2013, Freiburg im Breisgau9–12 April 2013
Masahiro Morii, Harvard WW, WZ, and ZZ
Diboson productionLO diagrams for WW, WZ, and ZZ production are
n Cross sections are calculated to NLOn Gluon-‐gluon enters at NLO. Less than 10% of the cross section
Allows access to triple gauge couplings (TGCs)n New physics may show up as anomalous (= non-‐SM) TGCs
Important background to Higgs and beyond-‐SM searchesn Precise knowledge of cross sections and kinematical distributions are important
2
W
q′
q
Z
W
TGC
s-‐channel
W /Z /γ
W /Z
W /Z q′ Z
q W
t-‐channel
W /Z
W /Zq′ Z
q W
u-‐channel
W /Z
W /Z
Masahiro Morii, Harvard WW, WZ, and ZZ
Leptonic final states (ℓ = e, μ and ν = νe, νμ, ντ)n High-‐pT isolated leptons + missing ET gives good S/B ratios
n Price: Br(W → ℓν) = 22%, Br(Z → ℓℓ) = 7%, Br(Z → νν) = 20%
Semi-‐leptonic final statesn Br(W → qq) = 68%, Br(Z → qq) = 70% ➔ More signal, but huge background
n b-‐jet identification can enhance Z → bb signal
Measurements
3
W W
+ ′ −
ν ν
W Z
+ ′ −
ν ′ +
Z
+
−
Z
′ −
′ +
Z
ν
ν
Z
−
+
W
+
ν
V
jet
jet
V
jet
jet
Z
+
−
Z
ν
ν
V
jet
jet
V = W or Z
Masahiro Morii, Harvard WW, WZ, and ZZ
References
4
CDF CDF note 10957 ZZ → ℓℓℓℓ, ℓℓνν [9.7 fb-1]
CDF note 10973 WW + WZ → ℓνjj [8.9 fb-1]
CDF note 10968 VV → MET + jj [9.1 fb-1]
CDF note 10864 ZW + ZZ → ℓℓjj [8.9 fb-1]
PRD 86 (2012) 031104 WZ → ℓνℓℓ [7.1 fb-1]
PRL 104 (2010) 201801 WW → ℓνℓν [3.6 fb-1]
DØ arXiv:1301.1243 WW → ℓνℓν [9.7 fb-1]
PLB 718 (2012) 451 TGC limits based on the following three analyses
PRL 108 (2012) 181803 WW + WZ → ℓνjj [4.3 fb-1]
PRD 85 (2012) 112005 WZ → ℓνℓℓ, ZZ → ℓℓνν [8.6 fb-1]
PRD 84 (2011) 011103 ZZ → ℓℓℓℓ [6.4 fb-1]
arXiv:1204.4496 WZ → ℓνbb [7.5 fb-1], ZZ → ννbb [8.4 fb-1], and ZZ → ℓℓbb [7.5 fb-1]
ATLAS ATLAS-CONF-2013-021 WZ → ℓνℓℓ [13 fb-1] 8 TeVATLAS-CONF-2013-020 ZZ → ℓℓℓℓ [20 fb-1] 8 TeVATLAS-CONF-2012-157 WW + WZ → ℓνjj [4.7 fb-1] 7 TeV
arXiv:1211.6096 ZZ → ℓℓℓℓ, ℓℓνν [4.6 fb-1] 7 TeV
arXiv:1210.2979 WW → ℓνℓν [4.6 fb-1] 7 TeV
EPJC 72 (2012) 2173 WZ → ℓνℓℓ [4.6 fb-1] 7 TeV
CMS CMS-PAS-FSQ-12-010 γγ → WW → eνµν [5.05 fb-1] 7 TeV
arXiv:1301.4698 WW → ℓνℓν [3.5 fb-1] and ZZ → ℓℓℓℓ [5.3 fb-1] 8 TeVJHEP 1301 (2013) 063 ZZ → ℓℓℓℓ [5.0 fb-1] 7 TeV
EPJC 73 (2013) 2283 WW + WZ → ℓνjj [5.0 fb-1] 7 TeV
CMS-PAS-SMP-12-005 WW → ℓνℓν [4.9 fb-1] 7 TeV
CMS-PAS-EWK-11-010 WZ → ℓνℓℓ [1.1 fb-1] 7 TeV
Masahiro Morii, Harvard WW, WZ, and ZZ
ZZ → ℓℓℓℓZZ → ℓℓℓℓ is the cleanest of all channelsn Two e+e− or μ+μ− pairs in loose Z mass windowsn Very small background from WZ and Z + jets
n CMS allows Z → τ+τ− for one Z
n
5
Subleading Lepton-Pair Mass [GeV]20 40 60 80 100 120 140 160 180 200 220
Lead
ing
Lep
ton
-Pair
Mass [
GeV
]
20
40
60
80
100
120
140
160
180
200
220
Data
lllll→ZZ
-1 L dt = 20 fb∫= 8 TeVs
PreliminaryATLAS
Figure 2: Invariant mass of the leading Z candidate versus the invariant mass of the subleading Z candidate. Theevents observed in the data are shown as solid circles and the signal prediction from simulation as pink boxes.The red box indicates the region defined by the ZZ fiducial cuts on the Z candidate masses. The blue dashed linesindicate the regions where one of the leptons pairs is in the mass window 66 < mll < 116 GeV. Contributions fromevents with one or both Z bosons outside of this mass window are also seen.
(lower pT) lepton pair. Events are required to contain two Z candidates with invariant masses satisfying66 < m`+`� < 116 GeV, resulting in 305 observed events.
4 Signal Acceptance
The ZZ selection requirements described above are applied to Monte Carlo simulation (PowhegBox andgg2zz), and corrections are applied to account for di↵erences in lepton reconstruction and identificatione�ciency, lepton energy scale and resolution, and trigger e�ciencies between data and simulation. Thecorrections are measured in data using samples of single Z ! `+`� events, supplemented by samplesof J/ ! e+e� events for electron e�ciency measurements at low pT. The overall e�ciency correctionfactor is 0.92±0.06 for the e+e�e+e� channel, 0.97±0.03 for the µ+µ�µ+µ� channel and 0.94±0.03 forthe e+e�µ+µ� channel, where the errors are systematic. A smearing and scale correction is added to themuon pT in the simulation [38] so that the Z ! µ+µ� invariant mass distribution in data is reproduced.Similarly, corrections are applied to the electron calorimeter energy resolution in the simulation and tothe electron calorimeter energy scale in the data [39]. A correction parameterized by the number of
5
ATLAS 8 TeV10 7 Summary
[GeV]2l2lm200 400 600 800
Even
ts /
20 G
eV
0
10
20
DATA ZZ WZ/Z + jets
CMS -1 = 8 TeV, L = 5.3 fbs
a)
[GeV]visτ2l2m
100 200 300 400 500 600
Even
ts /
25 G
eV
0
2
4
6 DATA ZZ WZ/Z + jets
CMS -1 = 8 TeV, L = 5.3 fbs
b)
[GeV]ll →Z1 m60 80 100 120
[GeV
]ll
→Z2
m
60
80
100
1202l2l → ZZ → pp
DATA 2eµ 4
µ 2e2
CMS -1 = 8 TeV, L = 5.3 fbs
c)
[GeV]ll →Z1 m60 80 100 120
[GeV
]vi
sττ
→Z2
m
20
40
60
80
100
120τ22l → ZZ → pp
DATA
CMS -1 = 8 TeV, L = 5.3 fbs
d)
Figure 2: Distributions for ZZ candidate events of (a) the four-lepton reconstructed mass for thesum of the 4e, 4µ, and 2e2µ channels and (b) the sum of the 2`2t channels. Points represent thedata, and shaded histograms represent the expected ZZ signal and the reducible background.The shapes of the signal and background are taken from the MC simulation, with each com-ponent normalized to the corresponding estimated value from Table 2. The distributions (c)and (d) demonstrate the relationship between the reconstructed Z1 and Z2 masses. Differentsymbols are used to present different decay channels.
CMS 8 TeV
ATLAS
-CO
NF-2013-020arX
iv:1301.4698
CMS 8 TeV
6
) [GeV/c]1
(lT
p0 50 100 150 200 250 300 350 400 450 5000
1
2
3
4
5
6
7DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(a)pT (l1).
) [GeV/c]2
(lT
p0 50 100 150 200 250 300 350 400 450 5000
1
2
3
4
5
6
DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(b)pT (l2).
) [GeV/c]3
(lT
p0 50 100 150 200 250 300 350 400 450 5000
2
4
6
8
10 DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(c)pT (l3).
) [GeV/c]4
(lT
p0 50 100 150 200 250 300 350 400 450 5000
1
2
3
4
5
6
7
8
DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(d)pT (l4).
R(ll)]Δmin[0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
2
2.5
3
3.5
4
4.5
DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(e)min(∆R(!!)).
R(ll)Δ0 0.5 1 1.5 2 2.5 3 3.5 40
1
2
3
4
5
6
7 Data
SM ZZ
Fakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(f)∆R(!!) of both lepton pair.
]2Z [GeV/cT
leading PllM20 40 60 80 100 120 140 160
]2Z
[GeV
/cT
sub
lead
ing
Pll
M
20
40
60
80
100
120
140
160
DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(g)M!! of the reconstructed Zs.
]2 [GeV/cllm0 50 100 150 200 2500
2
4
6
8
10
12 DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(h)m!! of both lepton pair.
(Z) [GeV/c]T
p0 50 100 150 200 250
Even
ts/(2
5 G
eV/c
)
0
1
2
3
4
5
6
7
8
9
10
DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(i)pT (!!) of both lepton pair.
]2 [GeV/c4lm100 200 300 400 500 600
Even
ts/(2
5 G
eV)
0
1
2
3
4
5
6
DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(j)M4!.
(ZZ) [GeV/c]T
p0 20 40 60 80 100 120 140 160 180 200
Even
ts/(2
0 G
eV)
0
1
2
3
4
5
6
7
8
9
10
DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(k)pT (ZZ).
Z decay angle-3 -2 -1 0 1 2 30
0.5
1
1.5
2
2.5
3
3.5
4
4.5
DataSM ZZFakes
CDF Run II Preliminary -1 L dt = 9.7 fb∫
(l)Z decay angle.
FIG. 1: Comparison of the MC simulation prediction and observation for some kinematic variables of the events inthe selected !!!′!′ collected sample.
CDF 9.7 fb-1
CD
F note 10957
Masahiro Morii, Harvard WW, WZ, and ZZ
ZZ → ℓℓννZZ → ℓℓνν looks for missing ET recoiling against a Z → ℓℓn Define “axial missing ET” =
n CDF and DØ use neural-‐net discriminators to separate signal from background
6
−ET ⋅pT pT
[GeV]missTAxial-E
-50 0 50 100 150
Even
ts /
5 G
eV
-1101
10210310410510610710810910
1010ATLAS
0 -1Ldt = 4.6 fb = 7 TeVs
DataZ+XW+X
*a/WaWTopWZ
WW-l+l-l+lAZZii
-l+lAZZTotal Uncertainty
ii-l+lAZZ
ATLAS 7 TeV
arXiv:1211.6096
8
The composition of the sample of events passing these requirements is summarized in TABLE III, including ex-pectations for other minor backgrounds. FIG. 3 shows some kinematic variable comparison between data and MCsimulation for events passing the Signal Region requirements.
) [GeV], Z axisTEφ Δ cos(TE--50 0 50 100 150 200
Even
ts /
5 G
eV
-210
-110
1
10
210
310
410
510
610
710
810
) [GeV], Z axisTEφ Δ cos(TE--50 0 50 100 150 200
Even
ts /
5 G
eV
-210
-110
1
10
210
310
410
510
610
710
810CDF Run II preliminary -1 L = 9.7 fb∫
ttγW
W+jetsDY
WZWW
ZZCDF Data
FIG. 2: E/TAx
distribution for the selected events (before
applying the E/TAx
requirement, comparing data to sim-ulation. The red line over impressed represents the ZZsignal shape, magnified ×5 with respect to the propernormalization, while all the solid process contributionare stacked.
∫
L = 9.7 fb−1
Process candidate eventsDY 317 ± 51.3tt 11.9 ± 2.2W+jets 69.5 ± 18.5Wγ 17.3 ± 2.2WW 114 ± 10.6WZ 37.5 ± 5.3Total Background 567 ± 24.4ZZ 63 ± 11Data 618
ZZ → !!νν Signal Region
TABLE III: Expected and observed number of eventspassing the kinematic requirements defining the SignalRegion.
1. Background modeling test
The modeling of the main background contribution in the defined Signal Region (Drell-Yan, WW ) have been testedin orthogonal sample of collected data with similar kinematic properties to the Signal Region. We test the Drell-Yanbackground modeling comparing data and simulation in a sample of events passing all the requirements applied forthe Signal Region definition but having E/T
Ax≤25 GeV, where the ZZ signal contribution is negligible. The WW
production modeling is tested comparing data to simulation in a sample of e±µ∓ events passing the same requirementsapplied for the Signal Region but 40 ≤ Meµ ≤ 140 GeV/c2, which has a negligible ZZ contribution and a small residualDrell-Yan contamination due to Z → ττ decays. The discrepancies between data and Monte Carlo simulation thataffects in particular the Drell-Yan modeling will be taken into account as a systematic uncertainties.
D. Neural Network separation
In order to improve the signal-to-background ratio further, we use an artificial neural network relying on thesimulated samples of signal and background events. This self-learning machine exploits kinematic information abouta given process (input variables) to produce an output value close to a target value (usually +1 for signal-like and-1 for background-like events). A NeuroBayes c© neural network (NN) [13] is trained using seven event kinematicvariables: the leading lepton transverse momentum (pT (#1)), the E/T significance (E/T /
√
∑
ET [20]), the dileptoninvariant mass (M!!), the dilepton system transverse momentum (p!!T ), the opening angles between the two leptons inthe transverse plane (∆φ(##)), the number of reconstructed jets in the events (Njets), and the angle in the detectortransverse plane between the E/T evaluated as the unbalance in the calorimeter towers and the P/T representing themissing transverse pT from the tracks reconstructed in the tracking system. These variables are the most sensitivefor signal-to-background separation since they exploit the unique features of ZZ production and are shown in Figure3, comparing data to simulations. Figure 4 shows the resulting NN output distributions for data and expected signaland background events in the Signal Region, in which ZZ signal events tend toward higher values and backgroundtoward lower values.
Exploiting the good separation of the signal from the background, we measure the ZZ cross section from a binnedmaximum likelihood fit of the NN output distribution. The likelihood function in the fit is the product of the Poisson
CD
F note 10957
CDF 9.7 fb-1
12
(GeV)’Tp10 20 30 40 50 60 70 80 90 100
Entri
es-110
1
10
210
310
410 )-e+Data (eZZ sig. Z bgd.WW bgd. Other bgd.
-1(a)D0 8.6 fb
(GeV)’Tp10 20 30 40 50 60 70 80 90 100
Entri
es
-110
1
10
210
310
410 )-µ+µData (
ZZ sig. Z bgd.
WW bgd. Other bgd.
-1(b)D0 8.6 fb
(GeV)’Tp10 20 30 40 50 60 70 80 90 100
Entri
es
0
10
20
30
40
50 )µData (e
ZZ sig. Z bgd.
WW bgd. Other bgd.
-1(c)D0 8.6 fb
(GeV)llM40 60 80 100 120 140 160 180 200 220 240
Entri
es
0
5
10
15
20
25
30
35 )-e+Data (eZZ sig. Z bgd.WW bgd. Other bgd.
-1(d)D0 8.6 fb
(GeV)llM40 60 80 100 120 140 160 180 200 220 240
Entri
es
0
5
10
15
20
25
30 )-µ+µData (
ZZ sig. Z bgd.
WW bgd. Other bgd.
-1(e)D0 8.6 fb
(GeV)llM40 60 80 100 120 140 160 180 200 220 240
Entri
es
0
10
20
30
40
50
60 )µData (e
ZZ sig. Z bgd.
WW bgd. Other bgd.
-1(f)D0 8.6 fb
Neural network output-3 -2 -1 0 1 2 3
Entri
es
0
5
10
15
20
25
30
35
40 )-e+Data (eZZ sig. Z bgd.WW bgd. Other bgd.
-1(g)D0 8.6 fb
Neural network output-3 -2 -1 0 1 2 3
Entri
es
0
5
10
15
20
25
30 )-µ+µData (
ZZ sig. Z bgd.
WW bgd. Other bgd.
-1(h)D0 8.6 fb
Neural network output-3 -2 -1 0 1 2 3
Entri
es
0
5
10
15
20
25
30
35
40
45 )µData (e
ZZ sig. Z bgd.
WW bgd. Other bgd.
-1(i)D0 8.6 fb
FIG. 6: (a-c) The p/′T distribution of the ZZ → !+!−νν candidate events before imposing the p/′T requirement. (d-f) The M!!
distribution of the ZZ → !+!−νν candidate events before imposing the M!! requirement. (g-i) The neural network outputdistribution of the accepted ZZ → !+!−νν candidate events. For the e±µ∓ channel, the neural network trained in the e+e−
channel is shown. The vertical dashed lines indicate the requirements on p/′T and M!!. The signal normalization is as describedin Section IV.
TABLE III: Table of predicted signal and background yields for the ZZ → e+e−νν signal and control regions. The systematicuncertainties are provided for the predictions.
Rejected by requirement onProcess Accepted p/′T a/′T Mll Extra lep. Charge Jets
Z → e+e− 0.6 ± 0.3 11666 ± 1665 0 ± 1 0.3 ± 0.2 3 ± 2 0.0 ± 0.0 0.1 ± 0.1
Z → τ+τ− 0.1 ± 0.1 8 ± 2 1.4 ± 0.2 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0WW → !+ν!−ν 35 ± 1 35 ± 1 1.7 ± 0.1 33 ± 1 9 ± 5 0.3 ± 0.1 0.1 ± 0.1WZ → !ν!+!− 2.3 ± 0.1 1.9 ± 0.1 0.2 ± 0.0 0.2 ± 0.1 14 ± 2 0.2 ± 0.1 0.0 ± 0.0W → eν 6 ± 2 13 ± 2 0.3 ± 0.1 5 ± 1 2 ± 1 4 ± 1 0.0 ± 0.0Wγ → eνγ 3.3 ± 0.3 5.5 ± 0.5 0.0 ± 0.1 2.8 ± 0.5 0.6 ± 0.5 3.3 ± 0.4 0.0 ± 0.0ZZ → !+!−!+!− 0.0 ± 0.0 0.1 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 1.3 ± 0.2 0.0 ± 0.0 0.0 ± 0.0tt → !+!−ννbb 1.0 ± 0.2 1.4 ± 0.2 0.4 ± 0.1 1.2 ± 0.1 7 ± 1 0.0 ± 0.0 0.2 ± 0.1Predicted background 48 ± 2 11749 ± 1668 4 ± 1 43 ± 2 37 ± 11 8 ± 1 0.4 ± 0.2
ZZ → !+!−νν 13.6 ± 0.4 7.4 ± 0.2 1.3 ± 0.1 0.6 ± 0.0 4 ± 2 0.2 ± 0.0 0.1 ± 0.0Predicted total 62 ± 3 11756 ± 1668 6 ± 1 43 ± 2 41 ± 13 8 ± 1 0.4 ± 0.2
Observed 61 10560 12 50 63 12 1
DØ 8.6 fb-1
eeννDØ 8.6 fb-1
µµνν
PR
D 85 (2012) 112005
Masahiro Morii, Harvard WW, WZ, and ZZ
ZZ → ℓℓℓℓ and ℓℓνν
Measured cross sections agreewith the SM predictionn NB: “total” cross section depends on the Z mass window
Measurement precisions arestatistics-‐limitedn Leading systematic errors is the leptonidentification efficiency
7
s L dt∫ Measured cross section (pb) Theory (pb)
CDF 1.96 TeV 9.7 fb−1 1.38 ± 0.19(stat)−0.19+0.20(sys) 1.4 ± 0.1 CDF note 10957
DØ 1.96 TeV 6.4–8.6 fb−1 1.40−0.37+0.43(stat) ± 0.14(sys) 1.4 ± 0.1 PRD 85 (2012) 112005
ATLAS 7 TeV 4.6 fb−1 6.7 ± 0.7(stat)−0.3+0.4(sys) ± 0.3(lumi) 5.89−0.18
+0.22 arXiv:1211.6096
CMS 7 TeV 5.0 fb−1 6.24−0.80+0.86(stat)−0.32
+0.41(sys) ± 0.14(lumi) 6.3 ± 0.4 EPJC 73 (2013) 2283
ATLAS 8 TeV 20 fb−1 7.1−0.4+0.5(stat) ± 0.3(sys) ± 0.2(lumi) 7.2−0.2
+0.3 ATLAS-CONF-2013-020
CMS 8 TeV 5.3 fb−1 8.4 ±1.0(stat) ± 0.7(sys) ± 0.4(lumi) 7.7 ± 0.4 arXiv:1301.4698
[TeV] s0 2 4 6 8 10 12 14
[pb]
ZZto
tal
σ
1
10
<116 GeV)ll) (66<mpZZ (p<116 GeV)llZZ (pp) (66<m
=8 TeV)sLHC Data 2012 (
=7 TeV)sLHC Data 2011 (
=1.96 TeV)sTevatron (
-1<116 GeV) L=20 fbll llll (66<m→ATLAS ZZ-1<120 GeV) L=5.3 fbll llll (60<m→CMS ZZ
-1<116 GeV) L=4.6 fbll) (66<mνν ll(ll/→ATLAS ZZ-1<120 GeV) L=5.0 fbll llll (60<m→CMS ZZ
-1<120 GeV) L=8.6 fbll) (60<mνν ll(ll/→D0 ZZ-1) (on-shell) L=6.0 fbνν ll(ll/→CDF ZZ
PreliminaryATLAS NLO QCD (MCFM, CT10.0)
Figure 4: Comparison of experimental measurements and theoretical predictions of the total ZZ produc-tion cross section as a function of centre-of-mass energy
ps. Shown are experimental measurements
from CDF [14] and D0 [15] in pp collisions at the Tevatron atp
s = 1.96 TeV, and experimental mea-surements from ATLAS [5] and CMS [6, 8] in pp collisions at the LHC at
ps = 7 TeV and
ps = 8 TeV.
The blue dashed line shows the theoretical prediction for the ZZ production cross section in pp collisions,calculated at NLO in QCD using MCFM with PDF set CT10. The solid red line shows the theoreticalprediction for the ZZ production cross section in pp collisions, calculated in the same way. The theo-retical curves are calculated using the natural width of the Z boson in the mass range 66 to 116 GeV.Atp
s = 8 TeV, the theoretical prediction using the zero-width approximation is 5% higher than theprediction using the natural width, restricted to the mass range 66 to 116 GeV.
12
Masahiro Morii, Harvard WW, WZ, and ZZ
[GeV]llM60 70 80 90 100 110 120
Even
ts /
2 G
eV
0
50
100
150
200
250
300
350 dataWZZZ
γW/Z+W+jetZ+jetTop
ATLAS Preliminary-1 L dt = 13 fb∫ = 8 TeV, s
[GeV]WZT
p0 50 100 150 200 250
Even
ts /
10 G
eV
020406080
100120140160180200220 data
WZZZ
γW/Z+W+jetZ+jetTop
ATLAS Preliminary-1 L dt = 13 fb∫ = 8 TeV, s
) [GeV]missT
(l,ETM20 40 60 80 100 120 140 160
Even
ts /
5 G
eV
0
20
40
60
80
100
120 dataWZZZ
γW/Z+W+jetZ+jetTop
ATLAS Preliminary-1 L dt = 13 fb∫ = 8 TeV, s
[GeV]WZM0 100 200 300 400 500 600 700 800
Even
ts /
20 G
eV
020406080
100120140160180 data
WZZZ
γW/Z+W+jetZ+jetTop
ATLAS Preliminary-1 L dt = 13 fb∫ = 8 TeV, s
Figure 4: Distributions of: the invariant mass of the Z candidate (top left), the WZ system transverse momentum(top right), the W transverse mass (bottom left) and the WZ invariant mass (bottom right). All these distributionsare shown after applying all of the WZ selection criteria in the four channels except the Z candidate mass. Thestatistical uncertainty is shown by shaded bands. For top (tt and single top) and Z+jets, the expected shape istaken from simulation but the normalization is taken from the data-driven estimates. The rightmost bins includeoverflow.
8
[GeV]llM60 70 80 90 100 110 120
Even
ts /
2 G
eV
0
50
100
150
200
250
300
350 dataWZZZ
γW/Z+W+jetZ+jetTop
ATLAS Preliminary-1 L dt = 13 fb∫ = 8 TeV, s
[GeV]WZT
p0 50 100 150 200 250
Even
ts /
10 G
eV
020406080
100120140160180200220 data
WZZZ
γW/Z+W+jetZ+jetTop
ATLAS Preliminary-1 L dt = 13 fb∫ = 8 TeV, s
) [GeV]missT
(l,ETM20 40 60 80 100 120 140 160
Even
ts /
5 G
eV
0
20
40
60
80
100
120 dataWZZZ
γW/Z+W+jetZ+jetTop
ATLAS Preliminary-1 L dt = 13 fb∫ = 8 TeV, s
[GeV]WZM0 100 200 300 400 500 600 700 800
Even
ts /
20 G
eV
020406080
100120140160180 data
WZZZ
γW/Z+W+jetZ+jetTop
ATLAS Preliminary-1 L dt = 13 fb∫ = 8 TeV, s
Figure 4: Distributions of: the invariant mass of the Z candidate (top left), the WZ system transverse momentum(top right), the W transverse mass (bottom left) and the WZ invariant mass (bottom right). All these distributionsare shown after applying all of the WZ selection criteria in the four channels except the Z candidate mass. Thestatistical uncertainty is shown by shaded bands. For top (tt and single top) and Z+jets, the expected shape istaken from simulation but the normalization is taken from the data-driven estimates. The rightmost bins includeoverflow.
8
WZ → ℓνℓℓWZ → ℓνℓℓ is quite clean: S/B ~ 4n e+e− or μ+μ− pair in a tight Z mass window, plus one isolated lepton and missing ET
n Loose cut on mT of W candidate
n Background from Z + jets and ZZ
Enough statistics to measure kinematicaldistributions, e.g. pT(Z) and m(WZ)
n ATLAS “unfolded” (= corrected for experimental resolutions) pT(Z) and m(WZ) distributions with 7 TeV data
8
ATLAS
-CO
NF-2013-021
11
Table 7 Normalized fiducial cross-sections and uncertainties in bins of pZT.
pZT [GeV] [0, 30] [30, 60] [60, 90] [90, 120] [120, 150] [150, 180] [180, 2000]
∆σfidWZ
(pZT)/σfidWZ
0.231 0.350 0.230 0.065 0.045 0.042 0.038Uncertainty 0.034 0.039 0.033 0.019 0.015 0.014 0.013
(a)
[GeV]TZp
0 30 60 90 120 150 180 2000Data
/MC
0.51
1.5
fid WZ
σ/fid W
ZσΔ
0.1
0.2
0.3
0.4
0.5
0.6ATLAS
= 7 TeV)sData 2011 (-1 L dt = 4.6 fb∫
Monte Carlo (MC@NLO)Data Total Uncertainty
(b)
[GeV]WZm170 270 405 2500Da
ta/M
C
0.51
1.5
fid WZ
σ/fid W
ZσΔ
0.2
0.4
0.6
0.8
1ATLAS
= 7 TeV)sData 2011 (-1 L dt = 4.6 fb∫
Monte Carlo (MC@NLO)Data Total Uncertainty
Fig. 7 Normalized fiducial cross-sections ∆σfidWZ
/σfidWZ
in binsof (a) pZT and (b) mWZ compared with the SM prediction.The total uncertainty contains statistical and systematic un-certainties added in quadrature.
for each source of the uncertainty, and combining theresulting changes in the unfolded spectrum. Because ofthe normalization, the results are affected only by theuncertainties that depend on pZT or mWZ . The stabilityof the unfolding procedure is tested in two ways: firstlyby comparing the unfolded spectra after two and af-ter three iterations, and secondly by checking that thetrue variable distribution is correctly reproduced froma simulated sample generated with non-zero anomalouscouplings.
7 Conclusion
Measurements of W±Z production in proton-protoncollisions at
√s = 7 TeV have been presented using a
Table 8 Normalized fiducial cross-sections and uncertaintiesin bins of mWZ .
mWZ [GeV] [170, 270] [270, 405] [405, 2500]
∆σfidWZ
(mWZ)/σfidWZ
0.568 0.283 0.149Uncertainty 0.038 0.030 0.027
data sample with an integrated luminosity of 4.6 fb−1,collected with the ATLAS detector at the LHC. Thecandidate W±Z events were selected in the fully lep-tonic final states with electrons, muons, and large miss-ing transverse momentum. In total, 317 candidates wereobserved with a background expectation of 68 ± 10events. The fiducial and total cross-sections are deter-mined to be
σfidWZ = 92+7
−6(stat.)± 4(syst.)± 2(lumi.) fb,
and
σtotWZ = 19.0+1.4
−1.3(stat.)± 0.9(syst.)± 0.4(lumi.) pb,
respectively, where the fiducial cross-section is definedby pµ,eT > 15GeV for the leptons from the decay ofthe Z bosons, pµ,eT > 20GeV for the leptons from thedecay of the W± bosons, |ηµ,e| < 2.5, pνT > 25GeV,|m"" − mZ| < 10GeV, MW
T > 20GeV, and ∆R > 0.3between the two leptons of all possible pairings of thethree leptons. These results are significantly more pre-cise than the earlier ATLAS measurement [6] whichthis paper supersedes. The total cross-section is con-sistent with the SM expectation of 17.6+1.1
−1.0 pb. Limitson anomalous triple gauge couplings have been derivedbased on the observed pZT distribution. The 95% confi-dence intervals are
∆gZ1 ∈ [−0.057, 0.093]
∆κZ ∈ [−0.37, 0.57]
λZ ∈ [−0.046, 0.047]
without a form factor. The limits are again more strin-gent than the earlier ATLAS measurement. Normalizedfiducial cross-sections have also been presented in binsof pZT and mWZ , and are in good agreement with SMpredictions.
EP
JC 72 (2012) 2173
ATLAS 7 TeVunfolded
ATLAS 8 TeV
ATLAS 8 TeV
9
(GeV)’TE0 20 40 60 80 100 120 140
Entri
es
0
10
20
30
40
50
60Data WZ sig.Z bgd.Other bgd.
-1(a) D0 8.6 fb
(GeV)llM40 60 80 100 120 140 160 180 200 220 240
Entri
es
0
10
20
30
40
50
60 Data WZ sig.Z bgd.Other bgd.
-1(b) D0 8.6 fb
(GeV)WTM
0 20 40 60 80 100 120 140 160 180 200
Entri
es
0
5
10
15
20
25 Data WZ sig.Z bgd.Other bgd.
-1(c) D0 8.6 fb
(GeV)T
Leading lepton p0 50 100 150 200 250
Entri
es
0
5
10
15
20
25
30
35 Data WZ sig.Z bgd.Other bgd.
-1(d) D0 8.6 fb
(GeV)T
Subleading lepton p0 20 40 60 80 100 120 140
Entri
es
0
5
10
15
20
25
30
35
40 Data WZ sig.Z bgd.Other bgd.
-1(e) D0 8.6 fb
(GeV)T
W daughter lepton p20 40 60 80 100 120 140 160 180 200
Entri
es
0
5
10
15
20
25
30
35 Data WZ sig.Z bgd.Other bgd.
-1(f) D0 8.6 fb
(GeV)llT
p0 50 100 150 200 250 300
Entri
es
0
5
10
15
20
25
30
35
40 Data WZ sig.Z bgd.Other bgd.
-1(g) D0 8.6 fb
(GeV)WT
p0 50 100 150 200 250 300
Entri
es
0
10
20
30
40
50Data WZ sig.Z bgd.Other bgd.
-1(h) D0 8.6 fb
FIG. 3: Kinematic distributions for the WZ → !ν!+!− signal candidates after combining the different sub-channels. Thefollowing variables are shown: (a) the E/ ′
T ; (b) the invariant mass of the Z → !+!− decay; (c) the W transverse mass; thetransverse momenta of the (d) leading and (e) subleading leptons from the Z → !+!− decay and (f) the charged lepton fromthe W decay; the transverse momenta of the reconstructed (g) Z → !+!− and (h) W → lν decays. The vertical dashed linesindicate the requirements on E/ ′
T and M!!. The signal normalization is as described in Section IV.
llTa
llLa
(2)T
p (1)T
p
llT
p
φΔ
t
Hadronic recoil
FIG. 4: Illustration of the decomposition of the dilepton pTinto aT and aL components.
allowed for the upward variation and protects againstelectrons for which the calorimeter has severely under-measured the energy. The amount by which, e.g., a!!T isreduced, is denoted aδT . These quantities are defined insuch a way that they always carry a negative sign.
B. Calorimeter recoil
Two estimates of the calorimeter recoil are made, fromthe reconstructed jets and from the reconstructed E/T .Jets are reconstructed using the D0 mid-point cone al-gorithm [29] with a cone size of ∆R = 0.5. They mustbe separated from the leptons by at least ∆R > 0.3 andsatisfy pT > 15 GeV. The pT , aT , and aL components
PR
D 85 (2012) 112005
DØ 8.6 fb-1
Masahiro Morii, Harvard WW, WZ, and ZZ
WZ → ℓνℓℓ
Measured cross sections agreewith the SM predictionn NB: “total” cross section depends on the Z mass window
Measurement precisions arestatistics-‐limitedn Exception: ATLAS 8 TeV, 13 m-‐1
n Leading systematic error is Z + jetsbackground estimate
9
s L dt∫ Measured cross section (pb) Theory (pb)
CDF 1.96 TeV 7.1 fb−1 3.93−0.53+0.60(stat)−0.46
+0.59(sys) 3.50 ± 0.21 PRD 86 (2012) 031104
DØ 1.96 TeV 8.6 fb−1 4.50 ± 0.61(stat)−0.25+0.16(sys) 3.21± 0.19 PRD 85 (2012) 112005
ATLAS 7 TeV 4.6 fb−1 19.0−1.3+1.4(stat) ± 0.9(sys) ± 0.4(lumi) 17.6−1.0
+1.1 EPJC 72 (2012) 2173
CMS 7 TeV 1.1 fb−1 17.0 ± 2.4(stat) ±1.1(sys) ±1.0(lumi) (19.8 ± 0.1) CMS-PAS-EWK-11-010
ATLAS 8 TeV 13 fb−1 20.3−0.7+0.8(stat)−1.1
+1.2(sys)−0.6+0.7(lumi) 20.3 ± 0.8 ATLAS-CONF-2013-021
[TeV] s0 2 4 6 8 10 12 14
[pb]
to
tal
WZ
σ
1
10 <116 GeV)ll
)(66<mpWZ (p<116 GeV)
llWZ (pp)(66<m
=8 TeV)sLHC Data 2012 (
=7 TeV)sLHC Data 2011 (
=1.96 TeV)sTevatron (
-1<116 GeV) L=13 fbllll (66<mν l→ATLAS WZ
-1<116 GeV) L=4.6 fbllll (66<mν l→ATLAS WZ
-1<120 GeV) L=8.6 fbllll (60<mν l→D0 WZ-1ll L=7.1 fbν l→CDF WZ
PreliminaryATLAS
NLO QCD (MCFM, CT10)
Figure 5: Measurements and theoretical predictions of the total W±Z production cross section as a functionof center-of-mass energy. Experimental measurements from CDF and D0 in proton antiproton collisions at theTevatron at
ps = 1.96 TeV, and experimental measurements from ATLAS in proton-proton collisions at the LHC
atp
s = 7 TeV andp
s = 8 TeV are shown. The blue dashed line shows the theoretical prediction for the W±Zproduction cross section in proton anti-proton collisions, calculated at NLO using MCFM with PDF set CT10. Thesolid red line shows the theoretical prediction for the W±Z production cross section in proton-proton collisions,calculated in the same way. The ATLAS results at 8 TeV define the total cross section with a Z boson with massbetween 66 GeV and 116 GeV. The results from CDF define the total cross section assuming zero-width for the Zboson and neglecting the �⇤ contribution. The results from D0 define the total cross section with a Z boson withmass between 60 GeV and 120 GeV.
12
Masahiro Morii, Harvard WW, WZ, and ZZ
WW → ℓνℓνWW → ℓνℓν needs a tight event selection to fight backgroundn Two isolated opposite-‐sign leptonsn Jet veto to suppress top background
▶ DØ uses 0-‐jet and 1-‐jet samplesn More kinematical cuts to suppress W/Z/γ* + jets
▶ ATLAS/CMS/CDF use a hard missing ET cut for e+e− and μ+μ− to suppress Z/γ* + jets
▶ DØ uses a BDT for e+e− or μ+μ−, mTmin and mT2 for e±μ∓
n ATLAS also unfolded the leading lepton pT distribution
10
8 6 Results
[GeV]maxT
p0 50 100 150
Even
ts /
5 G
eV
0
50
100
150
200 DATA WW VV Z + jets W + jets top
syst⊕ stat
CMS -1 = 8 TeV, L = 3.5 fbs
a)
[GeV]minT
p20 40 60 80
Even
ts /
5 G
eV
0
100
200
300
400 DATA WW VV Z + jets W + jets top
syst⊕ stat
CMS -1 = 8 TeV, L = 3.5 fbs
b)
[GeV]llT
p40 60 80 100 120
Even
ts /
5 G
eV0
100
200
DATA WW VV Z + jets W + jets top
syst⊕ stat
CMS -1 = 8 TeV, L = 3.5 fbs
c)
[GeV]llm0 50 100 150 200
Even
ts /
5 G
eV
0
50
100
150
DATA WW VV Z + jets W + jets top
syst⊕ stat
CMS -1 = 8 TeV, L = 3.5 fbs
d)
Figure 1: Distributions for W+W� candidate events of (a) the leading lepton transverse mo-mentum pmax
T , (b) the trailing lepton transverse momentum pminT , (c) the dilepton transverse
momentum p``T , and (d) the dilepton invariant mass m``. Points represent the data, and shadedhistograms represent the W+W� signal and the background processes. The last bin includesthe overflow. The W+W� signal is scaled to the measured cross section, and the backgroundprocesses are normalized to the corresponding estimated values in Table 1.
arXiv:1301.4698
CMS 8 TeV
1jet WW Discriminant0 0.2 0.4 0.6 0.8 1
Even
ts/0
.02
0
5
10
15
20
25
30
35
40
45 dataWWZ+jetsWZ,ZZW+jetsMultijetttbarBkg. syst.
TE + µ, e-1DØ, 9.7 fb
0jet WW Discriminant0 0.2 0.4 0.6 0.8 1
Even
ts/0
.02
0
20
40
60
80
100 dataWWZ+jetsWZ,ZZW+jetsMultijetttbarBkg. syst.
TE + µ, e-1DØ, 9.7 fb
DØ 9.7 fb-1
1-jetDØ 9.7 fb-1
0-jet
arXiv:1301.1243
12
[GeV]T
Leading lepton p40 60 80 100 120 140 160 180
Data
/MC
0.51
1.5
25 40 60 80 100 120 140 350
]-1 [G
eVT
/dp
WW
fid σ d×
WW
fid σ1/
0.005
0.01
0.015
0.02
0.025
0.03 ATLAS=7 TeV)sData 2011 (
-1 L dt = 4.6 fb∫
Monte Carlo (MC@NLO)DataStat. UncertaintyFull Uncertainty
FIG. 7: The normalized di↵erential WW fiducial cross section as a function of the leading lepton pT compared to the SMprediction.
Leading lepton pT [GeV] [25,40] [40,60] [60,80] [80,100] [100,120] [120,140] [140, 350]Weighted bin center [GeV] 33.6 50.2 70.2 89.1 107.1 127.5 180.41/�fid
WW ⇥ d�
fidWW /dpT [GeV�1] 2.0⇥ 10�2 2.1⇥ 10�2 8.2⇥ 10�3 2.7⇥ 10�3 2.2⇥ 10�3 9.5⇥ 10�4 6.2⇥ 10�5
Relative uncertainty 6.7% 4.8% 8.2% 17.0% 17.1% 25.5% 41.0%Correlation 1 �0.43 �0.33 �0.27 �0.27 �0.13 �0.29
1 �0.29 �0.29 �0.23 �0.30 �0.151 �0.01 �0.04 0.02 0.03
1 0.21 0.11 0.141 0.23 0.11
1 0.271
TABLE VII: Normalized fiducial cross section together with the overall uncertainty in bins of the leading lepton pT. Theweighted bin center is calculated as the cross-section-weighted average of the leading lepton pT in each bin derived frommc@nlo and gg2WW. The correlation coe�cients between di↵erent leading lepton pT bins are also shown. Only half of thesymmetric correlation matrix is presented.
A reweighting method is applied to SM WW eventsgenerated with mc@nlo and processed through the fulldetector simulation to obtain the leading lepton p
T
dis-tribution with anomalous couplings. The reweightingmethod uses an event weight to predict the rate withwhich a given event would be generated if anomalouscouplings were present. The event weight is the ratio ofthe squared matrix elements with and without anomalouscouplings i.e., |M|2/|M|2
SM
, where |M|2 is the matrixelement squared in the presence of anomalous couplingsand |M|2
SM
is the matrix element squared in the SM.The event generator bho [48] is used for the calculationof the two matrix elements. Generator-level comparisonsof WW production between mc@nlo and bho with allanomalous couplings set to zero are performed and con-sistent results are obtained. Samples with di↵erent setsof anomalous couplings are generated and the ratio of
the leading lepton pT
distribution to the SM predictionis parameterized as a function of the input anomalouscoupling parameters. This function is then used to inter-polate the leading lepton p
T
distribution for any givenanomalous couplings. To verify the reweighting method,the event weights for a given set of anomalous couplingsare calculated and applied to events generated with bho
assuming no anomalous couplings. The reweighted dis-tributions are compared to those predicted by the bho
generator, and good agreement is observed for the inclu-sive cross section and for the kinematic distributions asshown in Fig. 8(a).
Figure 8(b) compares the reconstructed leading lep-ton p
T
spectrum in data with that from the sum of ex-pected signal and background contributions. The pre-dicted leading lepton p
T
distributions for three di↵erentanomalous TGC values are also shown. Events at high
arXiv:1210.2979
ATLAS 7 TeVunfolded
Masahiro Morii, Harvard WW, WZ, and ZZ
WW → ℓνℓν
Cross sections at the LHC are slightly larger than the SM predictionn Significances are small (+1.4σ, +1.0σ, +1.7σ) but starting to draw attention
▶ Are the NLO calculations sufficiently precise?▶ Could this be a subtle sign of new physics?
Measurement precisions are systematics-‐limitedn Leading source of systematics is the jet veto
▶ Experimental: jet-‐energy scale and resolution affects jet pT threshold▶ Theoretical: number of jets in WW + jets
We still really want to see results from the full 8 TeV data
11
s L dt∫ Measured cross section (pb) Theory (pb)
CDF 1.96 TeV 3.6 fb−1 12.1± 0.9(stat)−1.4+1.6(sys) 11.7 ± 0.7 PRL 104 (2010) 201801
DØ 1.96 TeV 9.7 fb−1 11.6 ± 0.4(stat) ± 0.6(sys) 11.3 ± 0.7 arXiv:1301.1243
ATLAS 7 TeV 4.6 fb−1 51.9 ± 2.0(stat) ± 3.9(sys) ± 2.0(lumi) 44.7−1.9+2.1 arXiv:1210.2979
CMS 7 TeV 4.9 fb−1 52.4 ± 2.0(stat) ± 4.5(sys) ±1.2(lumi) 47.0 ± 2.0 CMS-PAS-SMP-12-005
CMS 8 TeV 3.5 fb−1 69.9 ± 2.8(stat) ± 5.6(sys) ± 3.1(lumi) 57.3−1.6+2.4 arXiv:1301.4698
Masahiro Morii, Harvard WW, WZ, and ZZ
WW/WZ → ℓνjjWW/WZ → ℓνjj is a tour de force of SM measurementsn Reconstruct a W candidate in one lepton + missing ETn Two jets with pT > 25–30 GeV
n Fit the di-‐jet invariant mass distribution
n Background must be exquisitely modeled
Signal is established by all experimentsn Precisions limited by both statistics and background modeling
12
=29.42χndof= 28
Data
W+X
Diboson
W/Z+jets
Top
QCD
]2Dijet-Mass [GeV/c50 100 150 200 250 300
2Ev
ents
/10
GeV
/c
0
0.2
0.4
0.6
0.8
1310×
-1CDF Run II Preliminary, L = 8.9 fb
]2Dijet-Mass [GeV/c50 100 150 200 250 300
2Ev
ents
/10
GeV
/c
-50
0
50
100
150Data-SM (no Diboson)
W+X
Diboson
FIG. 11: Fit to the dijet invariant mass distribution in the electron sample. All corrections
described in text have been applied here. Bottom figure shows data with all backgrounds (except
the diboson contribution) subtracted.
=69.42χndof= 56
Data
Diboson
W/Z+jets
Top
QCD
]2Dijet-Mass [GeV/c50 100 150 200 250 300
2Ev
ents
/5 G
eV/c
0100200300400500600700800
-1CDF Run II Preliminary, L = 8.9 fb
]2Dijet-Mass [GeV/c50 100 150 200 250 300
2Ev
ents
/5 G
eV/c
-50
0
50
100
150 Data-SM (no Diboson)
Diboson
FIG. 12: Fit to the dijet invariant mass distribution similar to [4]. The corrections described in
text have been applied here. Bottom figure shows data with all backgrounds (except the diboson
contribution) subtracted.
IX. ORTHOGONAL SAMPLES
Similar analyses are performed in two orthogonal samples. Same fit in invariant mass is
done as described in Sec. V and same kinematic cuts are applied as in Sec. II. The only
differences are in the objects selected.
20
=29.42χndof= 28
Data
W+X
Diboson
W/Z+jets
Top
QCD
]2Dijet-Mass [GeV/c50 100 150 200 250 300
2Ev
ents
/10
GeV
/c
0
0.2
0.4
0.6
0.8
1310×
-1CDF Run II Preliminary, L = 8.9 fb
]2Dijet-Mass [GeV/c50 100 150 200 250 300
2Ev
ents
/10
GeV
/c
-50
0
50
100
150Data-SM (no Diboson)
W+X
Diboson
FIG. 11: Fit to the dijet invariant mass distribution in the electron sample. All corrections
described in text have been applied here. Bottom figure shows data with all backgrounds (except
the diboson contribution) subtracted.
=69.42χndof= 56
Data
Diboson
W/Z+jets
Top
QCD
]2Dijet-Mass [GeV/c50 100 150 200 250 300
2Ev
ents
/5 G
eV/c
0100200300400500600700800
-1CDF Run II Preliminary, L = 8.9 fb
]2Dijet-Mass [GeV/c50 100 150 200 250 300
2Ev
ents
/5 G
eV/c
-50
0
50
100
150 Data-SM (no Diboson)
Diboson
FIG. 12: Fit to the dijet invariant mass distribution similar to [4]. The corrections described in
text have been applied here. Bottom figure shows data with all backgrounds (except the diboson
contribution) subtracted.
IX. ORTHOGONAL SAMPLES
Similar analyses are performed in two orthogonal samples. Same fit in invariant mass is
done as described in Sec. V and same kinematic cuts are applied as in Sec. II. The only
differences are in the objects selected.
20
CD
F note 10973
CDF 8.9 fb-14
(GeV)jjm50 100 150
Even
ts /
GeV
0
500
1000
1500
WW/WZW+jetstopQCDZ+jetsdata
TeV = 7s, -1fb dt = 5.0L∫CMS,
(a)
(GeV)jjm50 100 150
Even
ts /
GeV
0
50
100WW/WZdataUncertainty
TeV = 7s, -1fb dt = 5.0L∫CMS,
(b)
(GeV)jjm50 100 150
pull
-4
-2
0
2
4
TeV = 7s, -1fb dt = 5.0L∫CMS,
(c)
Figure 1: (a) Distribution of the dijet invariant mass in data, with the binning chosen based onthe resolution and fit projections of the relevant components overlaid. (b) The dijet invariantmass after subtraction of all components except the electroweak WW+WZ processes. The er-ror bars represent the statistical uncertainties and the hatched bands represent the systematicuncertainties. (c) The normalized residual or pull: (data � fit)/(fit uncertainty).
validate the fit procedure by performing pseudo-experiments. In each experiment, we generatethe mjj pseudo-data for the SM processes, taking into account the correlations between theyields, and then perform a fit to each pseudo-data mjj distribution. The results for both themuon and electron channels indicate that there is a small bias (-8.6% and -6.6%) in the WW+WZyield, corresponding to less than 0.4 standard deviations, and that the fit slightly overestimatesthe uncertainty on the yield. These effects are corrected for in the final result. The validationprocedure shows that biases in all background yields and errors are small. The fit results for thebackground components are statistically consistent with the expectations, with the exceptionof W+jets, where 11% fewer events for muons and 15% fewer events for electrons, compared tothe expectation, are observed. Overall, the approach produces a high quality model of the data(Fig. 1(a)), where the pull distribution is consistent with 0 (Fig. 1(c)), and allows us to extractthe diboson peak (Fig. 1(b)).
EP
JC 73 (2013) 2283
CMS 7 TeV
ATLAS
-CO
NF-2012-157
0 20 40 60 80 100 120 140 160 180 200
Entri
es /
5 G
eV
0
5000
10000
15000
20000
25000DataMultijet
+single-t ttW/Z + HFW/Z + jetsWW/WZ
-1L dt = 4.7 fb∫ = 7 TeVs
ATLAS Preliminary
+ 2 jetsν) µ l (e,→W
[GeV]T
Leading jet p0 20 40 60 80 100 120 140 160 180 200(D
ata-
MC
)/MC
-0.10.00.1
0 20 40 60 80 100
Entri
es /
5 G
eV
0
10000
20000
30000
40000
50000
60000
70000
80000 DataMultijet
+single-t ttW/Z + HFW/Z + jetsWW/WZ
-1L dt = 4.7 fb∫ = 7 TeVs
ATLAS Preliminary
+ 2 jetsν) µ l (e,→W
[GeV]T
SubLeading jet p0 20 40 60 80 100(D
ata-
MC
)/MC
-0.10.00.1
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Entri
es /
0.15
0
5000
10000
15000
20000
25000
30000
35000 DataMultijet
+single-t ttW/Z + HFW/Z + jetsWW/WZ
-1L dt = 4.7 fb∫ = 7 TeVs
ATLAS Preliminary
+ 2 jetsν) µ l (e,→W
R ΔDijet 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5(D
ata-
MC
)/MC
-0.10.00.1
50 100 150 200 250
Entri
es /
5 G
eV
0
2000
4000
6000
8000
10000
12000
14000
16000
18000 DataMultijet
+single-t ttW/Z + HFW/Z + jetsWW/WZ
-1L dt = 4.7 fb∫ = 7 TeVs
ATLAS Preliminary
+ 2 jetsν) µ l (e,→W
Dijet Mass [GeV]50 100 150 200 250(D
ata-
MC
)/MC
-0.10.00.1
Figure 2: Distributions of the leading (top left) and sub-leading (top right) jet transverse momenta, of their angulardistance ∆R (bottom left) and of the di-jet invariant mass distribution of reconstructed W/Z → j j candidates(bottom right). The distributions are for the electron and muon channels summed together. Filled circles show theexperimental data and the stacked histograms are SM predictions. The rightmost bins include overflow. In eachplot, the lower panel displays the fractional difference between the data and the MC expectation. The yellow bandsshow the JES systematic uncertainty, which dominates.
6
Dijet Mass [GeV]
ATLAS 7 TeV
10
(l) (GeV)T
p0 20 40 60 80 100 120 140
Even
ts /
(5 G
eV)
0
2000
4000
6000
8000
10000
12000
14000
16000 DataWVW/Z+LPW/Z+HFTopMultijetsUncert.
-1DØ 4.3 fb
) (GeV)1
(jetT
p0 20 40 60 80 100 120 140 160 180 200
Even
ts /
(5 G
eV)
0
2000
4000
6000
8000
10000
12000DataWVW/Z+LPW/Z+HFTopMultijetsUncert.
-1DØ 4.3 fb
) (GeV)2
(jetT
p0 20 40 60 80 100 120 140
Even
ts /
(5 G
eV)
05000
10000
1500020000250003000035000
4000045000 Data
WVW/Z+LPW/Z+HFTopMultijetsUncert.
-1DØ 4.3 fb
(GeV)TE0 20 40 60 80 100 120 140
Even
ts /
(5 G
eV)
02000400060008000
1000012000140001600018000 Data
WVW/Z+LPW/Z+HFTopMultijetsUncert.
-1DØ 4.3 fb
(GeV)jjM0 50 100 150 200 250 300
Even
ts /
(5 G
eV)
0
1000
2000
3000
4000
5000
6000DataWVW/Z+LPW/Z+HFTopMultijetsUncert.
-1DØ 4.3 fb
(GeV)νlTM
0 20 40 60 80 100 120 140 160 180 200
Even
ts /
(5 G
eV)
0
2000
4000
6000
8000
10000
12000 DataWVW/Z+LPW/Z+HFTopMultijetsUncert.
-1DØ 4.3 fb
(GeV)TH0 50 100 150 200 250 300
Even
ts /
(5 G
eV)
0100020003000400050006000700080009000
DataWVW/Z+LPW/Z+HFTopMultijetsUncert.
-1DØ 4.3 fb
(GeV)W)1
(dijet,jetrelT
p0 10 20 30 40 50 60 70 80 90 100
Even
ts /
(2 G
eV)
0
2000
4000
6000
8000
10000
12000 DataWVW/Z+LPW/Z+HFTopMultijetsUncert.
-1DØ 4.3 fb
FIG. 6: (color online) Distributions of the variables (next eight of fifteen) used as inputs to the RF classifier for electron andmuon channels combined, and before b-tagging. The signal and background predictions and the systematic uncertainty bandare evaluated after the fit of the total WV cross section in the RF output distribution. Definitions for each variable are providedin the text (LP denotes light partons such as u, d, s or gluon, and HF denotes heavy-flavor such as cc or bb).
PR
L 108 (2012) 181803
DØ 4.3 fb-1
Masahiro Morii, Harvard WW, WZ, and ZZ
Cross sections at the LHC
13
W Z WW Wt
[pb]
tota
lσ
1
10
210
310
410
510
-120 fb
-113 fb
-15.8 fb
-15.8 fb
-14.6 fb
-12.1 fb-14.6 fb
-14.6 fb-11.0 fb
-11.0 fb
-135 pb
-135 pb
tt t WZ ZZ
= 7 TeVsLHC pp Theory
)-1Data (L = 0.035 - 4.6 fb
= 8 TeVsLHC pp Theory
)-1Data (L = 5.8 - 20 fb
ATLAS PreliminaryATLAS PreliminaryATLAS Preliminary
[p
b]to
tσ
Prod
uctio
n C
ross
Sec
tion,
1
10
210
310
410
510
CMSNov 2012
W
1j≥
2j≥
3j≥
4j≥
Z
1j≥
2j≥
3j≥
4j≥
> 30 GeV jetTE
| < 2.4 jetη|
γW
> 15 GeVγ TE
,l) > 0.7γR(Δ
γZ WW+WZ WW
WZZZ
-136, 19 pb -15.0 fb -11.1 fb-15.0 fb -14.9 fb
-13.5 fb
-14.9 fb-15.3 fb
JHEP10(2011)132JHEP01(2012)010
CMS-PAS-SMP-12-011 (W/Z 8 TeV)
CMS EWK-11-009 CMS-PAS-EWK-11-010 (WZ)CMS-PAS-SMP-12-005 (WW7),
007(ZZ7), 013(WW8), 014(ZZ8), 015(WV)
syst) ⊕7 TeV CMS measurement (stat
syst) ⊕8 TeV CMS measurement (stat
7 TeV Theory prediction
8 TeV Theory prediction
W Z WW Wt
[pb]
tota
lσ
1
10
210
310
410
510
-120 fb
-113 fb
-15.8 fb
-15.8 fb
-14.6 fb
-12.1 fb-14.6 fb
-14.6 fb-11.0 fb
-11.0 fb
-135 pb
-135 pb
tt t WZ ZZ
= 7 TeVsLHC pp Theory
)-1Data (L = 0.035 - 4.6 fb
= 8 TeVsLHC pp Theory
)-1Data (L = 5.8 - 20 fb
ATLAS PreliminaryATLAS PreliminaryATLAS Preliminary
[p
b]to
tσ
Prod
uctio
n C
ross
Sec
tion,
1
10
210
310
410
510
CMSNov 2012
W
1j≥
2j≥
3j≥
4j≥
Z
1j≥
2j≥
3j≥
4j≥
> 30 GeV jetTE
| < 2.4 jetη|
γW
> 15 GeVγ TE
,l) > 0.7γR(Δ
γZ WW+WZ WW
WZZZ
-136, 19 pb -15.0 fb -11.1 fb-15.0 fb -14.9 fb
-13.5 fb
-14.9 fb-15.3 fb
JHEP10(2011)132JHEP01(2012)010
CMS-PAS-SMP-12-011 (W/Z 8 TeV)
CMS EWK-11-009 CMS-PAS-EWK-11-010 (WZ)CMS-PAS-SMP-12-005 (WW7),
007(ZZ7), 013(WW8), 014(ZZ8), 015(WV)
syst) ⊕7 TeV CMS measurement (stat
syst) ⊕8 TeV CMS measurement (stat
7 TeV Theory prediction
8 TeV Theory prediction
WW
WZ
ZZ
W Z WW Wt
[pb]
tota
lσ
1
10
210
310
410
510
-120 fb
-113 fb
-15.8 fb
-15.8 fb
-14.6 fb
-12.1 fb-14.6 fb
-14.6 fb-11.0 fb
-11.0 fb
-135 pb
-135 pb
tt t WZ ZZ
= 7 TeVsLHC pp Theory
)-1Data (L = 0.035 - 4.6 fb
= 8 TeVsLHC pp Theory
)-1Data (L = 5.8 - 20 fb
ATLAS PreliminaryATLAS PreliminaryATLAS Preliminary
ZZ
Masahiro Morii, Harvard WW, WZ, and ZZ
Triple Gauge CouplingsWWV (V = Z/γ) couplings ⬌ WW and WZ (also Wγ)
n 5 parameters:n Additional constraints may be imposed
ZZV (V = Z/γ) couplings ⬌ ZZ (also Zγ)
n 4 parameters:
Parameters in red (anomalous TGCs) are zero in the SM
14
W
q′
q
Z
W
TGC
Z /γ
W
W
LWWV
gWWV
= ig1V (Wµν
+W µV ν −Wµ+VνW
µν ) + iκVWµ+WνV
µν +iλV
mW2 Wλµ
+WνµV νλ
LZZV = −
eMZ
2 f4V (∂4
VV µβ )Zα (∂αZβ ) + f5V (∂σVσµ ) Z µβZβ
⎡⎣ ⎤⎦
Δg1
Z (≡ g1Z −1), Δκ Z (≡κ Z −1), Δκγ (≡κγ −1), λZ , λγ
f4Z , f4
γ , f5Z , f5
γ
W
q′
q
Z
W
TGC
W
W
Z
W
q′
q
Z
W
TGC
Z /γ
Z
Z
Equal coupling Δg1Z = 0, Δκ Z = Δκγ , and λZ = λγ
LEP scenario Δg1Z − Δκ Z = Δκγ tan2θW and λZ = λγ
HISZ scenario Δκ Z = Δg1Z (cos2θW − sin2θW ), Δκγ = 2Δg1
Z cos2θW and λZ = λγ
Masahiro Morii, Harvard WW, WZ, and ZZ
Triple Gauge CouplingsEffects of anomalous TGCs increase with n Increase sensitivity by binning in, or selecting the upper tail of n Observables: mZZ (for ZZ), pT of Z (for WZ or ZZ), pT of leading lepton (for WW)
n Also used: pT of di-‐jet in WW/WZ → ℓνjj
Extraction of TGC relies on NLO calculations: Powheg, MC@NLO, MCFMn LO-‐to-‐NLO correction is substantial at large
15
s s
7
include final states with hadronically decaying t leptons.
(GeV)2l2lm500 1000 1500
Even
ts/b
in
-110
1
10
DATAZZWZ/Z + jets
-1 = 7 TeV, L = 5.0 fbsCMS
= 0Z4f = 0.015Z
4f
Figure 2: Distribution of the four-lepton reconstructed mass for the sum of the 4e, 4µ, and the2e2µ channels. Points represent the data, and the shaded histograms represent the expected ZZsignal and the reducible background. The dashed and dotted histograms represent the resultsof the SHERPA simulation for the SM ( f Z
4 = 0) and in the presence of an ATGC ( f Z4 = 0.015),
while all the other anomalous couplings are set to zero.
The limits on ATGCs are calculated with the modified frequentist construction CLs [37–39]based on the shape of the four-lepton invariant mass distributions, including the 4e, 4µ, and2e2µ channels in the likelihood combination. Figure 2 presents the distribution of the four-lepton reconstructed mass for the sum of the 4e, 4µ, and 2e2µ channels. The dashed and dottedhistograms represent the results of the SHERPA simulation for the SM ( f Z
4 = 0) and in thepresence of an ATGC ( f Z
4 = 0.015), while all the other anomalous couplings are set to zero. Thepresence of ATGCs would be manifested in an increased yield of events at high four-leptonmasses. The invariant mass distributions are interpolated from SHERPA simulation for differentvalues of the anomalous couplings. For each distribution only one or two couplings are varied,while all others are set to zero. The fit is performed to find the maximum likelihood value andlimits are calculated. To avoid unitarity violation at energies above the scale L of new physics,the ATGCs are often modified with a form-factor parametrization of the type 1/(1 + s/L2)2,where
ps ⇡ m2`2` is the effective center-of-mass energy of the collision. However, no unitarity
violations occur in the sensitive region m2`2` . 1.5 TeV for bare anomalous couplings of order0.05 or smaller [40], so we calculate the limits without form-factor scaling. This choice has theadvantage of avoiding any bias from energy-dependence assumptions and is exact in the limitin which the scale of new physics is much larger than
ps.
Figure 3 presents the expected and observed two-dimensional exclusion limits at 95% confi-dence level (CL) on the anomalous neutral trilinear ZZZ and ZZg couplings. The green andyellow bands represent the one and two standard-deviation variations from the expected limit.The present limits are dominated by statistical uncertainties. Systematic uncertainties arisingfrom the uncertainty on the theoretical cross section, PDFs, detector efficiencies, and luminosityare introduced in the form of nuisance parameters with log-normal probability density func-tions. One-dimensional 95% CL limits for the f Z,g
4 and f Z,g5 anomalous coupling parameters
are measured to be
CMS ZZ 7 TeV
JHE
P 1301 (2013) 063
9
[GeV]ZT
p0 30 60 90 120 150 180 2000
Even
ts /
bin
0
20
40
60
80
100
120
140
160
DataSM WZBkg
stat + systσ = 0.05Zλ
= 0.57ZκΔ = 0.10Z
1gΔ
ATLAS
∫ -1Ldt = 4.6 fb = 7 TeVs
Fig. 4 Transverse momentum pZT of the Z boson inW±Z can-didate events. Data are shown together with expected back-ground and signal events, assuming the Standard Model. Ex-pected events in the case of anomalous TGC without formfactor are also shown for illustration. The last bin is short-ened for display purposes.
squared of the W±Z system, approaches infinity. Toachieve this, an arbitrary form factor may be intro-duced [32]. Here the dipole form factor adopted is
α(s) =α0
(1 + s/Λ2)2(8)
where α stands for ∆gZ1 , ∆κZ , or λZ , α0 is the valueof the anomalous coupling at low energy, and Λ is thecut-off scale, the scale at which new physics enters. Theresults are reported both with and without this formfactor.
Since an enhancement in the cross-section due to ananomalous coupling would grow with s, measurementsensitivity to anomalous TGCs is enhanced by binningthe data in a kinematic variable related to s. The trans-verse momentum pZT of the Z boson provides a natu-ral choice for such binning as it is strongly correlatedwith s and can be directly reconstructed from the mea-sured lepton momenta with good precision. The dataare therefore divided into six bins in pZT of width 30GeVfollowed by a wide bin that includes 180–2000GeV.
MC@NLO [11] is used to generateW±Z events withnon-SM TGC. The generator computes, for each event,a set of weights that can be used to reweight the fullsample to any chosen set of anomalous couplings. Thisfunctionality is used to express the predicted signalyields in each bin of pZT as a function of the anoma-lous couplings. Figure 4 shows the pZT distribution ofthe selected events together with the SM prediction.Also shown for illustration are predictions with non-zero anomalous couplings without form factor: each cou-pling is increased to the expected 99% confidence-levelupper limit while keeping the other two couplings atthe SM value. For this plot the 99%, rather than 95%,
Table 6 Expected and observed 95% confidence intervals onthe anomalous couplings ∆gZ1 , ∆κZ , and λZ . The expectedintervals assume the Standard Model values for the couplings.
Observed Observed ExpectedΛ = 2 TeV no form factor no form factor
∆gZ1 [−0.074, 0.133] [−0.057, 0.093] [−0.046, 0.080]∆κZ [−0.42, 0.69] [−0.37, 0.57] [−0.33, 0.47]λZ [−0.064, 0.066] [−0.046, 0.047] [−0.041, 0.040]
confidence-level upper limits are used to accentuate dif-ferences in shape. As expected, the largest deviationsfrom the SM are in the last bin of pZT, while the devia-tions in the lower-pZT bins depend on which coupling isvaried.
Frequentist confidence intervals are obtained on theanomalous couplings by forming a profile likelihood testincorporating the observed number of candidate eventsin each pZT bin, the expected signal as a function ofthe anomalous couplings and the estimated number ofbackground events [33]. The systematic uncertaintiesare included in the likelihood function as nuisance pa-rameters with correlated Gaussian constraints. A pointin the anomalous TGC space is accepted (rejected) atthe 95% confidence level if less (more) than 95% of ran-domly generated pseudo-experiments exhibit a value ofthe profile likelihood ratio larger than that observed indata.
Table 6 summarizes the observed 95% confidenceintervals on the anomalous couplings ∆gZ1 , ∆κZ , andλZ , with the cut-off scale Λ = 2 TeV and without theform factor. The limits on each anomalous TGC param-eter are obtained with the other two anomalous TGCparameters set to zero. The expected intervals in Ta-ble 6 are medians of the 95% confidence-level upper andlower limits obtained in pseudo-experiments that as-sume the SM coupling. The widths of the expected andobserved confidence intervals are dominated by statisti-cal uncertainty. Figure 5 compares the observed limitswith the Tevatron results [34, 35].
The 95% confidence regions are shown as contourson the (∆gZ1 ,∆κZ), (∆gZ1 ,λZ), and (∆κZ ,λZ) planesin Figure 6. In each plot the remaining parameter is setto the SM value. The limits were derived with no formfactor.
6.3 Normalized Fiducial Cross-Sections
The effective Lagrangian adopted in the TGC analysisin Section 6.2 allows us to probe non-SM physics withlittle model dependence. An alternative approach is tomeasure kinematic distributions, such as the pZT spec-trum, that could be compared with model-dependent
ATLAS WZ 7 TeVE
PJC
72 (2012) 2173
7
0 50 100 150 200 250 300 350 400 450 500-210
-110
1
10
210
-W+SM W
Background
Data
= 0.16ZλΔ
= 0.34Z1
gΔ
= 0.72γκΔ
]-1 [GeV cT
Lepton p
-1Ev
ents
/ 20
GeV
c
FIG. 2: Leading-lepton pT distribution for data compared tothe SM expectation. Also shown is how the expectation wouldbe modified by anomalous couplings near the observed limits.
TABLE II: Expected and observed limits on anomalousTGCs. For each coupling limit set, the two other couplingsare fixed at their SM values. Values of the couplings outside ofthe given observed range are excluded at the 95% confidencelevel.
Λ (TeV) λZ ∆gZ1 ∆κγ
Expected 1.5 (-0.05,0.07) (-0.09,0.17) (-0.23,0.31)Observed 1.5 (-0.16,0.16) (-0.24,0.34) (-0.63,0.72)Expected 2.0 (-0.05,0.06) (-0.08,0.15) (-0.20,0.27)Observed 2.0 (-0.14,0.15) (-0.22,0.30) (-0.57,0.65)
observed 95% confidence limits, shown in Table II, areweaker than expected. The probability of observing theselimits in the presence of only standard model W+W−
production ranges from 7.1% to 7.6% depending on thecoupling constants (λZ , gZ
1 ,κγ).In summary, the W+W− production cross section has
been measured in pp collisions at√
s = 1.96 TeV fromreconstructed events in the dilepton final state using alikelihood ratio formed from matrix-element-based eventprobabilities. This result is the most precise measure-ment at this energy with an overall uncertainty of lessthan 15%. The same event sample is also used to per-form the most sensitive probe to date at this energy ofanomalous WWZ and WWγ couplings. The leading-lepton pT distribution of the sample is found to be inmoderate agreement with the SM expectation and usedto place limits on anomalous triple gauge couplings.
We thank the Fermilab staff and the technical staffsof the participating institutions for their vital contribu-tions. This work was supported by the U.S. Departmentof Energy and National Science Foundation; the ItalianIstituto Nazionale di Fisica Nucleare; the Ministry ofEducation, Culture, Sports, Science and Technology ofJapan; the Natural Sciences and Engineering Research
Council of Canada; the National Science Council of theRepublic of China; the Swiss National Science Founda-tion; the A.P. Sloan Foundation; the Bundesministeriumfur Bildung und Forschung, Germany; the World ClassUniversity Program, the National Research Foundationof Korea; the Science and Technology Facilities Coun-cil and the Royal Society, UK; the Institut National dePhysique Nucleaire et Physique des Particules/CNRS;the Russian Foundation for Basic Research; the Minis-terio de Ciencia e Innovacion, and Programa Consolider-Ingenio 2010, Spain; the Slovak R&D Agency; and theAcademy of Finland.
[1] J. M. Campbell and R. K. Ellis, Phys. Rev. D 60, 113006(1999).
[2] S. Frixione and B. R. Webber, J. High Energy Phys. 06(2002) 029.
[3] J. Ellison and J. Wudka, Ann. Rev. Nucl. Part. Sci. 48,33 (1998).
[4] F. Abe et al. (CDF Collaboration), Phys. Rev. Lett. 78,4536 (1997).
[5] D. E. Acosta et al. (CDF Collaboration), Phys. Rev. Lett.94, 211801 (2005).
[6] V. M. Abazov et al. (D0 Collaboration), Phys. Rev. Lett.94, 151801 (2005) [Erratum-ibid. 100, 139901 (2008)].
[7] V. M. Abazov et al. (D0 Collaboration), Phys. Rev. Lett.(to be published) [arXiv:0904.0673].
[8] J. Abdallah et al. (DELPHI Collaboration), Eur. Phys.J. C 54, 345 (2008); S. Schael et al. (ALEPH Collabora-tion), Phys. Lett. B 614, 7 (2005); P. Achard et al. (L3Collaboration), Phys. Lett. B 586, 151 (2004); G. Abbi-endi et al. (OPAL Collaboration), Eur. Phys. J. C 33, 463(2004); V. M. Abazov et al. (D0 Collaboration), Phys.Rev. D 74, 057101 (2006); T. Aaltonen et al. (CDF Col-laboration), Phys. Rev. D 76, 111103 (2007).
[9] K. Hagiwara, S. Ishihara, R. Szalapski, and D. Zeppen-feld, Phys. Rev. D 48, 2182 (1993).
[10] A. Abulencia et al. (CDF Collaboration), J. Phys. G 34,2457 (2007).
[11] T. Aaltonen et al. (CDF Collaboration), Phys. Rev. Lett.102, 021802 (2009).
[12] T. Sjostrand, S. Mrenna, and P. Skands, J. High EnergyPhys. 05 (2006) 026.
[13] U. Baur, T. Han, and J. Ohnemus, Phys. Rev. D 57,2823 (1998).
[14] R. Brun, R. Hagelberg, M. Hansroul, and J. C. Lassalle,version 3.15, CERN-DD-78-2-REV.
[15] S. Moch and P. Uwer, Nucl. Phys. Proc. Suppl. 183, 75(2008).
[16] C. Anastasiou, L. J. Dixon, K. Melnikov, and F. Petriello,Phys. Rev. D 69, 094008 (2004).
[17] D. Acosta et al., Nucl. Instrum. Methods Phys. Res.,Sect. A 494, 57 (2002).
[18] F. James and M. Roos, Comput. Phys. Commun. 10, 343(1975).
CDF WW 3.6 fb-1
PR
L 104 (2010) 201801 7
(GeV)jjT
p0 50 100 150 200 250 300
Even
ts / B
in210
310
410DataWVW/Z+l.f.W/Z+h.f.TopMultijetsUncert.
= 0.1,λ = 0.2γκΔ
-1DØ, L = 4.3 fb(a)
(GeV)llT
p0 20 40 60 80 100 120140 160 180200
Even
ts / B
in
05
101520253035 Data
WZDYOther bgd.Uncert.
= -0.05,λ = -0.06Z
1 gΔ
-1DØ, L = 8.6 fb(b)
FIG. 1: (color online) (a) The pjjT distribution summed over electron and muon channels from WW +WZ → !νjj (l = µ, e)production for data and SM MC predictions (“l.f.” denotes light partons such as u, d, s or gluon, and “h.f.” denotes heavy-flavor such as c or b). Also shown are expected distributions for an ATGC model with ∆κγ = 0.2, and λ = 0.1. (b) The pllTdistribution summed over eee, eµµ, µee and µµµ channels from WZ → !ν!! production for data, SM MC predictions and forATGC model with λ = −0.05 and ∆gZ1 = −0.06.
γκ Δ-0.4 -0.2 0 0.2 0.4
λ
-0.15-0.1
-0.050
0.050.1
0.150.2
0.250.3 -1DØ, L = 4.3 fb
(a)
LEP parameterization
Standard ModelMinimum95% Contour68% Contour
γκ Δ-0.4 -0.2 0 0.2 0.4
Z 1 gΔ
-0.15-0.1
-0.050
0.050.1
0.150.2
0.250.3
0.35-1DØ, L = 4.3 fb
(b)
LEP parameterization
Standard ModelMinimum95% Contour68% Contour
λ-0.1 -0.05 0 0.05 0.1
Z 1 gΔ
-0.15-0.1
-0.050
0.050.1
0.150.2
0.250.3
-1DØ, L = 4.3 fb(c)
LEP parameterization
Standard ModelMinimum95% Contour68% Contour
γκ Δ-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3
λ
-0.1
-0.05
0
0.05
0.1
0.15
0.2-1DØ, L = 4.3 fb
(d)
Equal couplings parameterization
Standard ModelMinimum95% Contour68% Contour
FIG. 2: WW +WZ → !νjj (l = µ, e). The 68% and 95% C.L. two-parameter limits on the γWW/ZWW coupling parametersassuming the LEP (a, b, c) and equal couplings parameterization (d) with Λ = 2 TeV. Black circles indicate the most probablevalues of an ATGCs from the two-parameter fit.
of systematic uncertainties on separate samples and sub-channels due to the same uncertainty are assumed to be100% correlated but different uncertainties are assumedto be uncorrelated.
The 68% and 95% C.L. limits on ATGCs from the4.3 fb−1 analysis of WW + WZ → !νjj final states inthe two-parameter space are shown in Fig. 2. The limitsfrom the 8.6 fb−1 analysis of WZ → !ν!! final states
are presented only in the λ − ∆gZ1 space as shown inFig. 3, because WZ production is weakly sensitive to∆κγ via the relation given by Eq. (3). The 95% C.L.one-parameter limits, obtained from single parameter fitswith all other parameters fixed to their SM values arepresented in Table I.
The resulting 68% and 95% C.L. one-parameter lim-its from the combined fit of !νγ, !ν!ν, !νjj, and !ν!!
DØ ℓνjj 4.3 fb-1
PLB
718 (2012) 451
s
Masahiro Morii, Harvard WW, WZ, and ZZ
TGC results
TGCs consistent with the SMn Four of the WWZ and WWγ couplings are constrained to O(0.05)▶ Caveat: LEP scenario is used▶ Δκγ remains less precise
n ZZZ and ZZγ couplings are constrained by the LHC results to O(0.01)
8 TeV data not included yet
16
WWZ couplings WWγ couplings
ZZZ and ZZγ couplings
Masahiro Morii, Harvard WW, WZ, and ZZ
Does it make sense?Shouldn’t the WW “excess” show up as anomalous TGCs?
n TGC sensitivity is concentrated in the highest bins of leading lepton pT
n Excesses are mostly at low pT where anomalous TGCs don’t contribute
If the excess is real, it’s not a kind of physics described with aTGCsn i.e. not a heavy new particle in s-‐channel loop diagrams
17
arXiv:1210.2979
ATLAS 7 TeV
5
(GeV)Tmax
p0 20 40 60 80 100 120 140 160 180 200
even
ts/b
in
0
20
40
60
80
100
120
140
160
180 Data WW WZ + ZZ Top Fakes Z+jets
syst.) ⊕ (stats.σ
CMS preliminary 2012-1 = 7 TeV, L = 4.92 fbs
(GeV)Tmax
p0 20 40 60 80 100 120 140 160 180 200
Rat
io
0
1
2 (GeV)
Tminp
0 20 40 60 80 100 120 140 160
even
ts/b
in
0
50
100
150
200
250
300
350
400 Data WW WZ + ZZ Top Fakes Z+jets
syst.) ⊕ (stats.σ
CMS preliminary 2012-1 = 7 TeV, L = 4.92 fbs
(GeV)Tmin
p0 20 40 60 80 100 120 140 160
Rat
io
0
1
2
(GeV)llT
p40 60 80 100 120 140 160 180 200
even
ts/b
in
0
50
100
150
200
250 Data WW WZ + ZZ Top Fakes Z+jets
syst.) ⊕ (stats.σ
CMS preliminary 2012-1 = 7 TeV, L = 4.92 fbs
(GeV)llT
p40 60 80 100 120 140 160 180 200
Rat
io
0
1
2
(GeV)invM0 20 40 60 80 100 120 140 160 180 200
even
ts/b
in
0
20
40
60
80
100
120
140 Data WW WZ + ZZ Top Fakes Z+jets
syst.) ⊕ (stats.σ
CMS preliminary 2012-1 = 7 TeV, L = 4.92 fbs
(GeV)invM0 20 40 60 80 100 120 140 160 180 200
Rat
io
0
1
2
Figure 1: Distributions of the leading lepton transverse momentum (pTmax), the trailing lep-ton transverse momentum (pTmin), and the dilepton transverse momentum (p``T ) and invariantmass (M``) at the final selection level, reweighted to the data-driven estimates. All four chan-nels (ee, µµ, eµ and µe) are combined, and the uncertainty band corresponds to the statisticaland systematic uncertainties on the predicted yield.
5 Efficiencies and Systematic Uncertainties
The signal efficiency is estimated using simulation, considering both the qq ! W+W� andgg ! W+W� prcosses. Residual discrepancies in the lepton reconstruction and identificationefficiency between data and simulation are corrected by determining data-to-simulation scalefactors measured using Z/g⇤! `+`� events in the Z peak region [25], recorded with dedicatedunbiased triggers. These factors depend on the lepton pT and |h| and are within four percent
CM
S-PA
S-S
MP
-12-005
CMS 7 TeV
Masahiro Morii, Harvard WW, WZ, and ZZ
SummaryWW, WZ, and ZZ measurements continue to improven CDF/DØ results with full Run-‐II datan ATLAS/CMS results with (full or partial) 8 TeV data
Data are largely consistent with the SM predictionn WW cross sections at the LHC slightly higher than expected
▶ This has spurred theoretical investigations▶ Jet veto is the leading source of experimental/theoretical uncertainties▶ Analyze full 8 TeV data!
TGCs show no deviation from the SMn Four (out of 5) WWZ and WWγ couplings constrained at ~0.05 level
▶ Δκγ remains less precise, but improving
n All ZZZ and ZZγ couplings constrained at ~0.01 leveln 8 TeV data have yet to be used ➔ Expect improvements soon
18
Backup Slides
Masahiro Morii, Harvard WW, WZ, and ZZ
Two-‐photon WW productionNew measurement from CMS of γγ → WWn Final state is e±μ∓ with no other tracks from the event vertex
n pT(eμ) > 30 GeV suppresses ττ background
n 2 events observed. 2.2 ± 0.5 signal and 0.84 ± 0.13 (stat.) background expected
Limits are set on anomalous quartic gauge couplings (aQGCs)
20
14 8 Results
Correcting for efficiency, acceptance, and backgrounds, the result interpreted as a cross sectiontimes branching fraction is:
s(pp ! p(⇤)W+W�p(⇤) ! p(⇤)µ±e⌥p(⇤)) = 2.1+3.1�1.9 fb,
with a significance of 1.1s. With statistical uncertainties only, the resulting value of the crosssection times branching fraction is 2.1+3.0
�1.9(stat.) fb. The SM prediction is 3.8 ± 0.9 fb, includingthe uncertainty on the contribution of proton dissociation.
The pT(µ±e⌥) distribution for events with zero extra tracks is shown in Figure 12. In the AQGCsearch region pT(µ±e⌥) > 100 GeV, zero events are observed in data, consistent with the Stan-dard Model expectation of 0.14, dominated by pp ! p(⇤)W+W�p(⇤).
) [GeV]µ(eT
p0 50 100 150 200 250 300
Even
ts/3
0 G
eV
0
5
10
15
20
25
30
=500GeV)Λ (a0W=2E-4, aCW=0, -W+ W→ γγ
=500GeV)Λ (a0W=-2E-4, aCW=-8E-4, -W+ W→ γγ
Data -τ+τDrell-Yan -W+Inclusive W -W+Diffractive W
-τ+τ → γγElastic -τ+τ → γγInelastic
tt W+jets
(SM)-W+ W→ γγ
-1=7 TeV, L=5.05 fbsCMS Preliminary 2011,
vertex)µN(extra tracks, e0 2 4 6 8 10 12 14
Even
ts
0
5
10
15
20
25
=500GeV)Λ (a0W=2E-4, aCW=0, -W+ W→ γγ
=500GeV)Λ (a0W=-2E-4, aCW=-8E-4, -W+ W→ γγ
Data -τ+τDrell-Yan -W+Inclusive W -W+Diffractive W
-τ+τ → γγElastic -τ+τ → γγInelastic
tt W+jets
(SM)-W+ W→ γγ
-1=7 TeV, L=5.05 fbsCMS Preliminary 2011,
Figure 12: Full pT(µ±e⌥) distribution for events with 0 extra tracks (left) and multiplicity of extratracks for events with pT(µ±e⌥) > 100 GeV (right). The backgrounds (solid histograms) are stackedwith statistical uncertainties indicated by the shaded region, the signal histogram (open histogram) isstacked on top of the backgrounds. The expected signal is shown for the SM gg ! W+W� signal (solidlines), and for two representative values of the anomalous couplings aW
0 /L2 and aWC /L2 (dotted and
dashed lines).
We find the selection efficiency does not vary strongly between the SM and AQGC samplestested within the acceptance (Table 5), and therefore set an upper limit on the partial cross sec-tion times branching fraction for gg ! W+W� ! µ±e⌥ with pT(µ, e) > 20 GeV, |h(µ, e)| < 2.4,and pT(µ±e⌥) > 100 GeV. We treat the residual SM pp ! p(⇤)W+W�p(⇤) signal as a back-ground, resulting in a total of 0.14 ± 0.02 expected events, and include an additional systematicuncertainty of 10% based on the maximum relative variation of the efficiency between the SMand samples generated with two values of the anomalous couplings.
We find an upper limit of 3 events, corresponding to a 95% CL upper limit on the partial crosssection times branching fraction with the selection cuts pT(µ, e) > 20 GeV, |h(µ, e)| < 2.4, andpT(µ±e⌥) > 100 GeV:
s(pp ! p(⇤)W+W�p(⇤) ! p(⇤)µ±e⌥p(⇤)) < 1.9 fb.
In Figure 13, the excluded value is compared to the predicted cross section for non-zero valuesof aW
0 /L2 and aWC /L2 within the defined acceptance, scaled to include the contribution from
proton dissociation. The theoretical prediction is determined from the fully simulated samples,
9
Selection step Signal e ⇥ A Visible cross section (fb) Events in dataTrigger and preselection 28.5% 1.4 9086m(µ±e⌥) > 20 GeV 28.0% 1.4 8200Muon ID and Electron ID 22.6% 1.1 1222µ±e⌥ vertex with 0 extra tracks 13.7% 0.7 6pT(µ±e⌥) > 30 GeV 10.6% 0.5 2
Table 2: Signal efficiency ⇥ acceptance and number of events selected in data at each stage of theselection. The preselection corresponds to requiring a reconstructed muon and electron of oppositecharge, each having pT > 15 GeV and |h| < 2.4, matched to a common primary vertex with less than 15additional tracks.
) [GeV]µ(eT
p0 50 100 150 200 250 300
Even
ts/6
GeV
-210
-110
1
10
210
310Data -τ+τDrell-Yan
-W+Inclusive W -W+Diffractive W-τ+τ → γγElastic -τ+τ → γγInelastic
tt W+jets (SM)-W+ W→ γγ
-1=7 TeV, L=5.05 fbsCMS Preliminary 2011,
Figure 7: µ±e⌥ pair transverse momentum. The events plotted are required to pass the trigger andpreselection requirements, and the lepton identification. The shaded bands indicate the statistical un-certainty on the background estimation.
to the other backgrounds, without accounting for any survival probabilities or overlap with theinclusive W+W� sample. To study the W+jets backgrounds, where the contribution is mainlyfrom fake leptons or non-prompt leptons in jets, we select a control sample of events withpT(µ±e⌥) > 30 GeV and at least one of the two leptons failing the nominal offline identification.This sample is then normalized to the simulation in the high-multiplicity (more than 6 extratracks) region, and used to estimate the W+jets background in the signal and inclusive W+W�
control regions. Figure 8 shows the distribution of the number of extra tracks for the W+W�
region with pT(µ±e⌥) > 30 GeV. In Figure 9, the invariant mass and acoplanarity of the eventswith 1-6 extra tracks are plotted. In general the data is consistent with the sum of simulatedbackgrounds.
The corresponding extra track multiplicity distribution for the Drell-Yan t+t�-dominated re-gion with pT(µ±e⌥) < 30 GeV is shown in Figure 8. We find a deficit in data compared tosimulation in the 1-6 tracks region. In the t+t� sample with zero extra tracks we find 4events in data, compared to an MC background expectation of 2.5 events, plus 0.9 events ofgg ! W+W� signal. The expected contribution to the background from gg ! t+t� is ap-proximately 0.7 events. The invariant mass and acoplanarity distributions are plotted in Fig-ure 10.
Table 3 summarizes the observed and expected background event yields for the three orthogo-nal control regions. Due to tracks from pile-up vertices being wrongly associated to the µ±e⌥
CMS 7 TeV
CM
S-PA
S-FS
Q-12-010CMS 7 TeV
No othertracks
15
with one anomalous quartic coupling parameter fixed to zero, and a second order polynomialinterpolation of the other. With a dipole form factor of L = 500 GeV, the limits obtained on theAQGC parameters are:
�0.00017 < aW0 /L2 < 0.00017 GeV�2 (aW
C /L2 = 0, L = 500 GeV),�0.0006 < aW
C /L2 < 0.0006 GeV�2 (aW0 /L2 = 0, L = 500 GeV),
which are approximately two orders of magnitude more stringent than the limits obtained atLEP [7, 10, 12].
We perform a similar procedure to scan the two dimensional space of aW0 /L2 and aW
C /L2, usinga large number of samples generated with a fast simulation of the CMS detector [44]. For eachpoint the cross section and the number of events passing generator selection requirements isused to calculate the observed cross section times branching fraction. This procedure onlydepends on the generator level prediction, with the fast simulation used to confirm that thesignal efficiency of all trigger, reconstruction, and analysis selections, relative to the acceptanceis flat across AQGC sample space. The result is shown in Figure 14, where the area outside theellipse corresponds to values of the anomalous couplings that would result in a partial crosssection times branching fraction above 1.9 fb, including the form factor with L = 500 GeV.The result of Ref. [7], obtained from a maximum likelihood fit to a combination of WWg andWW ! gg channels in e+e� collisions, is shown in the inset.
We also obtain the corresponding limits without form factors. In this case the cross section isdominated by the region of high energy gg interactions, above the unitarity bound. This leadsto limits on the anomalous couplings much smaller than in the scenario with form factors:
�2.80 ⇥ 10�6 < aW0 /L2 < 2.80 ⇥ 10�6 GeV�2 (aW
C /L2 = 0, no form factor),�1.02 ⇥ 10�5 < aW
C /L2 < 1.02 ⇥ 10�5 GeV�2 (aW0 /L2 = 0, no form factor),
]-2 [GeV2Λa0W/-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25
-310×
BF
[fb]
× σ
0
0.5
1
1.5
2
2.5
-1=7 TeV, L=5.05 fbsCMS Preliminary 2011,
= 500 GeVΛ
) > 20 GeVµ(e,T
p
)| < 2.4µ(e,η|
e) > 100 GeVµ(T
p
]-2 [GeV2ΛaCW/-0.0015 -0.001 -0.0005 0 0.0005 0.001 0.0015
BF
[fb]
× σ
0
0.5
1
1.5
2
2.5
-1=7 TeV, L=5.05 fbsCMS Preliminary 2011,
= 500 GeVΛ
) > 20 GeVµ(e,T
p
)| < 2.4µ(e,η|
e) > 100 GeVµ(T
p
Figure 13: Expected partial cross section times branching fraction with L = 500 GeV as a functionof aW
0 /L2 with aWC /L2 fixed to 0 (left), and aW
C /L2 with aW0 /L2 fixed to 0 (right). The prediction
(solid line) and excluded value (dashed line) are both shown for pT(µ, e) > 20 GeV, |h(µ, e)| < 2.4,pT(µ±e⌥) > 100 GeV. The prediction is rescaled to include the contribution from proton dissociation.
Top Related