- Email: [email protected]
Recent Electroweak Results from ATLAS
Available online at www.sciencedirect.com
Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246 www.elsevier.com/locate/nppp
Recent Electroweak Results from ATLAS Richard Batleya , on behalf of the ATLAS Collaboration a Department of Physics, University of Cambridge Cavendish Laboratory, J J Thomson Avenue, Cambridge CB3 0HE, UK.
Abstract √ √ Recent electroweak physics results obtained using proton-proton collision data at s = 7 TeV and s = 8 TeV recorded by the ATLAS experiment at the LHC are reviewed. The topics covered include: measurements of diboson cross sections, a measurement of the branching ratio for the rare decay Z → 4, and studies of vector boson fusion and vector boson scattering via measurements of the electroweak components of Z j j and W ± W ± j j production. The constraints on anomalous triple and quartic gauge couplings resulting from these measurements are also summarised. Keywords: LHC, ATLAS, electroweak 1. Introduction Electroweak interactions in the Standard Model (SM) derive from a local SU(2)×U(1) gauge symmetry. The expected triple (TGC) and quartic (QGC) gauge boson couplings can be tested at the LHC through studies of multiboson (W, Z, γ) production or through studies of vector boson fusion (VBF) and vector boson scattering (VBS) processes. Studies of VBS processes are also of interest as they can potentially test whether the single Higgs scalar boson of the SM is indeed solely responsible for unitarising the amplitude for longitudinal W boson scattering, WL WL → WL WL . Electroweak physics results from proton-proton collision data recorded by the ATLAS experiment [1] during √ √ the 2011 ( s = 7 TeV) and 2012 ( s = 8 TeV) runs of the LHC [2] are reviewed below. Emphasis is placed on the most recent results: a preliminary measurement of √ the W + W − cross section at s = 8 TeV; a measurement of the branching ratio for the rare decay Z → 4; and studies of VBF and VBS processes via measurements of electroweak Z j j and W ± W ± j j production. 2. W +W − production At the LHC, W + W − production occurs dominantly via the process qq → W + W − (see Fig. 1) and is sensitive to the WWZ and WWγ TGCs. Additional contributions from gg → W + W − or gg → H → W + W − contribute about 2% and 7%, respectively, to the total W + W − cross section. The W + W − cross section measured using the ATLAS 7 TeV data set [3], 51.9 ± 2.0(stat) ± http://dx.doi.org/10.1016/j.nuclphysbps.2015.01.051 2405-6014/© 2015 Published by Elsevier B.V.
3.9(stat) ± 2.0(lumi) pb, is in agreement with the predicted SM cross section of 44.7 ± 2.1 pb computed at next-to-leading order (NLO) QCD using MCFM [4]. A preliminary√ ATLAS measurement of the W + W − cross section at s = 8 TeV has now also been obtained [5].
q¯
W
q¯
W
q¯
q
Z/γ ∗
W
q
TGC vertex
W
Figure 1: Leading order Feynman diagrams for the process qq → W + W −
Candidate W + W − events in the ATLAS 8 TeV data set are selected by requiring two isolated, high pT , oppositely charged electrons or muons (e+ e− , μ+ μ− or e± μ∓ ) (leading lepton pT > 25 GeV, sub-leading pT > 20 GeV), accompanied by large missing transverse enmiss ergy: ET,rel > 45(15) GeV and pmiss > 45(20) GeV, T where the numbers in brackets correspond to the mixedmiss = ETmiss × sin |Δφ| flavour e± μ∓ channel. Here, ET,rel where Δφ is the azimuthal angle between the ETmiss vector and the closest selected jet or lepton in the event; miss rather than ETmiss reduces the impact of misusing ET,rel calibrations or fluctuations in the measurement of jet or lepton energies. The quantity ETmiss is obtained primarily from the calorimeter energies and reconstructed muons, while pmiss is obtained by summing over reconstructed T tracks. In addition, a requirement |Δφmiss | < 0.3(0.6) is placed on the difference in azimuthal directions of
242
Events
10000
s = 8 TeV, ∫ Ldt = 20.3 fb
600
ATLAS Preliminary
s = 8 TeV, ∫ Ldt = 20.3 fb-1
e± ν μ ν channel
Data Top Wjets stat. unc.
WW (× 1.21) Zjets other diboson stat. ⊕ syst. unc.
500 400 300 200 100 50
ATLAS Preliminary
100
150
200
250
300
Data -1
mll [GeV]
WW MC Top MC
e± ν μ ν channel ±
12000
700
±
the ETmiss and pmiss vectors. Events are rejected if they T contain addition electrons or muons with pT > 7 GeV, additional jets with pT > 25 GeV (“jet veto”), or, for the same-flavour (e+ e− , μ+ μ− ) channels, if the invariant mass of the dilepton pair lies within 15 GeV of the Z boson mass. The jet veto requirement is effective at reducing the irreducible background from top quark production, as shown in Fig. 2. Using the ATLAS 8 TeV data set, 6636 events are selected in the zero-jet bin, mostly (5067 events) in the e± μ∓ channel.
Events / 10 GeV
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
Zjets MC
Figure 3: Lepton pair invariant mass distribution for selected W + W − events in
Wjets MC
8000
the e± μ∓ channel [5]
other diboson MC stat. unc.
6000
3. Diboson production and anomalous TGCs
4000 2000 0
1
2
3
4
5
6
7
Jet multiplicity
Figure 2: Jet multiplicity distribution in W + W − events for the e± μ∓ channel, without imposing the jet veto requirement (which selects only the first bin) [5]
Fig. 3 shows the reconstructed dilepton mass distribution for the mixed-flavour (e± μ∓ ) channel. The expectation for the W + W − signal is obtained from the POWHEG MC [6] for the dominant qq → W + W − contribution, and from the GG2WW MC [7] for the gg → W + W − contribution. The expected background from top quark production and from (W/Z)+jets production is estimated from data, while the background from diboson production (other than W + W − ) is estimated from MC. In Fig. 3, the W + W − signal expectation has been scaled by a factor of 1.21 to obtain agreement with data. √ The W + W − total cross section at s = 8 TeV is +5.0 +2.2 measured to be 71.4+1.2 −1.2 (stat)−4.4 (syst)−2.1 (lumi) pb, and is about 2.1σ above the SM expectation of 58.7+3.0 −2.7 pb determined using MCFM for the qq → W + W − and gg → W + W − processes and using [8] for the gg → H → W + W − process. Various contributions neglected here may serve to increase the predicted cross section, including NNLO QCD corrections and a k-factor (possibly as large as a factor 2-3) for the gg → W + W − process. Also, the MCFM prediction is obtained using CT10 [9] parton distribution functions (PDFs); using alternative PDFs such as ATLAS-epWZ12 [10] may increase the predicted cross section by as much as 2.9 pb.
Cross section measurements have been published using the 7 TeV ATLAS data set for various diboson production channels: ZZ [11], WZ [12], WW [3], Zγ [13] and Wγ [13]. In all cases, the measured total cross sections are in agreement with the corresponding SM prediction computed at NLO QCD. The observed event rates at high lepton, photon or Z boson pT are used to place constraints on possible anomalous TGC contributions within a model-independent effective Lagrangian framework. The resulting constraints on the anomalous γ,Z coupling parameters f4,5 and hγ,Z 3,4 for the neutral ZZZ, γZZ vertices (forbidden in the SM) improve on those obtained by the combined LEP experiments by a large factor (typically 5-10). The sensitivities to the anomalous coupling parameters λγ,Z , Δκγ,Z , gZ1 for the charged vertices WWZ, WWγ have not yet surpassed those from LEP, though this should happen once analysis of the 8 TeV data is complete. At present, at 8 TeV, besides the W + W − cross section above, preliminary cross section measurements are available for ZZ production in the + − + − ( = e, μ) channel [14], and for WZ production [15]. 4. The rare decay Z → 4 The 4 mass spectrum obtained in observations of the Higgs boson resonance H → ZZ ∗ → 4 [16] shows a second, nearby resonant peak at the Z mass due to Z → 4 decay. In the SM, the Z → 4 resonance arises through s-channel qq → 4 diagrams such as that in Fig. 4(a), where a lepton from the Z decay radiates a virtual Z boson or photon which undergoes internal conversion
243
to a charged lepton pair. The contribution to 4 production in the Z resonance region from t-channel diagrams such as Fig. 4(b) is small, about (3-4)%, while the process gg → 4 contributes only about 0.1%. Unlike the decay H → 4, which involves only a virtual Z boson, the virtual photon component of the dominant s-channel Z → 4 process leads to a final state dominated by the presence of a low mass e+ e− or μ+ μ− pair. To measure the branching ratio BR(Z → 4) as inclusively as possible and with greater statistical precision, a looser cut, m > 5 GeV, than in the Higgs analysis is placed on the invariant mass of all same-flavour opposite-charge lepton pairs (e+ e− , μ+ μ− ) in the 4 system [17]. The minimum pT requirement on electrons and muons is lowered to 7 GeV and 4 GeV, respectively. q
−
120 ATLAS s = 7 TeV, 4.5 fb-1
100
s = 8 TeV, 20.3 fb-1
80
Data Z 10xBkg σstat+syst Z→4l
60 40 20 0 75
80
85
90
95
100
105
m4l [GeV] Figure 5: The invariant mass of the 4 system for Z → 4 candidate events, without imposing the requirement 80 < m4 < 100 GeV [17]
− Z ∗ /γ ∗
q
Events / 3 GeV
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
+ Z/γ ∗
Z ∗ /γ ∗ q¯
+
−
Z ∗ /γ ∗
−
+
q¯
+
Figure 4: Examples of (a) s-channel and (b) t-channel leading order Feynman diagrams contributing to qq → 4 production
Candidate Z → 4 decays are selected in the fourlepton invariant mass range 80 < m4 < 100 GeV, in the e+ e− e+ e− , μ+ μ− μ+ μ− and e+ e− μ+ μ− channels. A total of 152 events is selected for the 7 TeV and 8 TeV data sets and for all channels combined. The m4 distribution for the selected events is shown in Fig. 5. The total background from Z+jets and top quark production, estimated using data-driven methods, and from diboson production, estimated using MC samples, is about 1.0% of selected events. Interference between the s-channel and t-channel contributions is studied at leading order using CALCHEP [18], and found to be small, about 0.2%. The resonant s-channel contribution can therefore effectively be isolated by subtracting the expected fraction of events, fnr , due to the non-resonant t-channel and gg → 4 processes. The Z → 4 branching ratio is then determined by normalising to the observed rate of Z → μ+ μ− events in the same data sets: N 4 2μ , N 2μ 4 where N 4 and N 2μ are the number of backgroundsubtracted Z → 4 and Z → μ+ μ− events observed in the data, 4 and 2μ are the overall selection efficiencies for Z → 4 and Z → μ+ μ− events, respectively, and BR(Z → μ+ μ− ) = (3.366 ± 0.007)% [19]. Normalising to the Z → μ+ μ− data cancels the luminosity BR(Z → 4) = BR(Z → μ+ μ− )(1 − fnr )
uncertainty, and partially cancels the theoretical uncertainties. The branching ratio is determined for the phase space region defined by the requirements 80 < m4 < 100 GeV and m > 5 GeV for all e+ e− and μ+ μ− pairs in the 4 system. In this region, the Z → 4 branching ratio is measured to be (3.20 ± 0.25 ± 0.13) × 10−6 , in agreement with the SM prediction from POWHEG of (3.33 ± 0.01) × 10−6 . Interference between the schannel and t-channel contributions (0.2%) is included in the systematic error on the measured branching ratio. 5. Vector boson fusion: Z j j production The VBF production, qq → qqZ, of a Z boson via the fusion of two W bosons, illustrated in Fig. 6(a), gives rise to two final state jets from the scattered quarks which tend to be produced in the forward directions, at large rapidities. The VBF process is accompanied by other electroweak Z j j diagrams, and is dominated by strong (QCD) Z j j production processes such as the diagram shown in Fig. 6(b). The two jets in the Z j j final state tend to be produced with higher pT in the electroweak case, and tend to have larger invariant mass, m j j , and larger rapidity separation, |Δy j j |. Furthermore, the lack of any colour exchange between the scattered quarks in the electroweak VBF process results, on average, in less additional jet activity in the region between the two forward jets than for the QCD contributions. The selection of candidate Z j j events [20] requires at least two jets, with pT > 55(45) GeV for the leading (sub-leading) jet. The Z boson is reconstructed via the decays Z → e+ e− and Z → μ+ μ− , where the decay leptons are required to have pT > 25 GeV and in-
244
q
−
g
−
μ ,e
W+
jet
q
q
−
−
μ ,e
Z
Z
W−
+
+
μ ,e
1 dσ σ dNgap
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
+
0.6
Data (2012) Sherpa Zjj (QCD + EW)
0.5
Sherpa Zjj (QCD) Powheg Zjj (QCD + EW)
+
μ ,e
Powheg Zjj (QCD)
0.4 q
q
q
g
0.3
ATLAS
∫ L dt = 20.3 fb
-1
Figure 6: Examples of (a) electroweak VBF and (b) QCD Feynman diagrams
s = 8 TeV high-mass region
0.2
contributing to Z j j production
106
ATLAS
baseline region -1 ∫ L dt = 20.3 fb s = 8 TeV
105 104
Data (2012) QCD Zjj EW Zjj VZ Top Multijets
1
1.5 1 0.5 0
1
2
3
4
5
6
7
8
gap
Njet
103
gap Figure 8: The unfolded Njet distribution for the Z j j “high mass” region. The
102
expected distribution neglecting the electroweak Z j j component is given by the horizontal dashed lines [20]
10 Prediction Data
1.5
0.5
Powheg Data
Nobs/200 GeV
variant mass m in the range 81 < m < 101 GeV. The distribution of the invariant mass m j j is plotted in Fig. 7, and shows the enhanced electroweak component at high m j j , but also that the QCD component remains dominant even at high mass.
Sherpa Data
0.1
2 1.5 1 0
500
1000
1500
2000
2500 3000 mjj [GeV]
Figure 7: The m j j distribution in Z j j candidate events, before imposing any gap
requirements on m j j or Njet [20]
The “search” region is defined by the additional regap gap quirements m j j > 250 GeV and Njet = 0, where Njet is the number of additional reconstructed jets with pT > 25 GeV in the rapidity interval between the two leading jets. The “high mass” region is defined by increasing the jet-jet mass cut to m j j > 1 TeV and removing the gap Njet requirement. The electroweak component is expected to be about 4.0% (12%) of selected events for the “search” (“high mass”) regions, while backgrounds from diboson and top quark production are estimated to be small, about 1.3% (3%). Background-subtracted, unfolded differential distributions are obtained for a vagap riety of kinematic variables; the unfolded Njet distribution for events in the “high mass” region is shown in gap Fig. 8, for example. The Njet distribution is reasonably well modelled by both the (LO) SHERPA [21] and (NLO) POWHEG MCs. The reduced jet activity expected for the electroweak component is visible in Fig. 8 from a comparison of the solid and dashed distributions.
To extract the Z j j electroweak component (Z j j − EW), a template fit is made to the m j j distribution for selected events in the “search” region. The template for the Z j j − EW signal is obtained from the SHERPA MC, while the background template is obtained from the SHERPA MC for the QCD Z j j process, plus small diboson and top quark contributions from other MC samples. To account for possible mismodelling, the background template is constrained using a data “control” region obgap tained by reversing the jet veto (Njet ≥ 1). The fit to the m j j distribution is shown in Fig. 9, and gives a Z j j−EW cross section of 54.7 ± 4.6(stat)+9.8 −10.4 (syst) ± 1.5(lumi) fb, in agreement with the SM expectation computed using POWHEG. The background-only hypothesis is excluded at over 5σ, and the measurement thus represents an observation of the electroweak Z j j process. The subset of “search” region events with m j j > 1 TeV (900 events, with an expected Z j j − EW contribution of about 260 events) is used to constrain anomalous contributions to the WWZ triple gauge vertex. Limits are placed on the anomalous coupling parameters gZ1 and λZ which are less restrictive than those obtained to date from WZ production, but which are complementary in that the two W bosons involved in the Z j j VBF process are spacelike, not timelike.
245
10
103
ATLAS L dt = 20.3 fb-1 s = 8 TeV search region
102 10 Data (2012) Background
BKG BKG + EW Data Data
1
Background + EW Zjj
1.5 1 0.5 1 0.5 0
constrained unconstrained
500
1000 1500 2000 2500 3000 3500 mjj [GeV]
Figure 9: Results of the fit to the m j j distribution in Z j j events used to extract the electroweak Z j j component. The horizontal dashed lines in the lower panel show the effect of removing the “control” region constraint on the background template [20]
6. Vector boson scattering: W ±W ± j j production Examples of electroweak VBS contributions to the process qq → WW j j, involving WWVV (V = W, Z) QGCs or Higgs exchange, are shown in Fig. 10. Similarly to the case of Z j j production above, the VBS processes are accompanied by other electroweak contributions and are dominated by strong (QCD) contributions. Again, the VBS diagrams for WW j j production tend to produce two high pT , forward jets, and the VBS contribution can be enhanced by requiring large invariant mass, m j j , or large rapidity separation, |Δy j j |, of the two forward jets. Unlike Z j j production, and since at LO the two W bosons produced in gg → WW j j processes must have opposite electric charge, the QCD contribution to WW j j production can be reduced by a large factor by requiring that the two W bosons have the same charge.
q
q
q
q
W+
W+
q
q
q
102
10
ATLAS 20.3 fb-1, s = 8 TeV
Data 2012 Syst. Uncertainty W±W±jj Electroweak W±W±jj Strong Prompt Conversions Other non-prompt
1
10-1
5
Data/Bkg Bkg Uncertainty (Sig+Bkg)/Bkg
0 200 400 600 800 1000 1200 1400 1600 1800 2000 mjj [GeV]
Figure 11: The distribution of the invariant mass m j j of the jet-jet system in W ± W ± j j events. The vertical dashed line indicates the position of the cut m j j > 500 GeV [22]
W+
q
data [22] by requiring two jets with pT > 30 GeV, with jet-jet invariant mass m j j > 500 GeV. In addition, events are required to contain exactly two same-sign electrons or muons (e± e± , μ± μ± or e± μ± ), each with pT > 25 GeV, and to contain large missing transverse energy ETmiss > 40 GeV. To reduce backgrounds from diboson and top quark production, events with additional leptons (electrons with pT > 7 GeV or muons with pT > 6 GeV), or with reconstructed b-quark jets (with pT > 30 GeV and |η| < 2.5) are rejected. To reduce the background from (Z → e+ e− )+jets production, where the charge of one of the electrons is misreconstructed, events with e± e± mass within 10 GeV of the Z boson mass are removed. The m j j distribution for W ± W ± j j candidates for all channels combined is shown in Fig. 11. The signal expectation is evaluated using events generated with SHERPA and normalised to the cross section computed using POWHEG. Electroweak W ± W ± j j production becomes the dominant expected contribution at high values of m j j . The m j j > 500 GeV requirement leaves a sample of 50 events in the 8 TeV data, with an estimated irreducible background of about 20 events. The background-only hypothesis is excluded at 4.5σ, compared with an expected exclusion significance of 3.4σ.
Events/50 GeV
∫
4
Data/Background
Nobs / 250 GeV
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
W+
Figure 10: Examples of electroweak VBS Feynman diagrams contributing to W ± W ± j j production
Same-sign W ± W ± j j events are selected in the 8 TeV
Fig. 12 shows the |Δy j j | distribution for the sample of 50 selected W ± W ± j j candidates. The electroweak WW j j component is extracted using the “VBS” region, defined by imposing the additional requirement |Δy j j | > 2.4, leaving a data sample of 34 events. Af-
246
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
Events
ter subtracting the estimated irreducible background of about 15 events, and subtracting also the strong WW j j contribution of about 1.3 events, the electroweak WW j j cross section (including interference between the electroweak and strong WW j j contributions) is measured to be 1.3 ± 0.4(stat) ± 0.2(syst) fb, in good agreement with the SM expectation of 0.95 ± 0.06 fb computed using POWHEG. The background-only hypothesis is excluded with a significance of 3.6σ, compared with an expected significance of 2.8σ, and as such the measurement provides the first evidence for electroweak WW j j production. 30 25
ATLAS 20.3 fb-1, s = 8 TeV mjj > 500 GeV
Data 2012 Syst. Uncertainty W±W±jj Electroweak W±W±jj Strong Prompt Conversions Other non-prompt
20 15 10
yet the 8 TeV data. Even so, experimental constraints on anomalous ZZZ and ZZγ couplings are now dominated by those obtained at the LHC, while the constraints obtained on anomalous WWZ and WWγ couplings are already approaching the combined constraints from the LEP experiments. The branching ratio for the rare decay Z → 4 has been determined with a precision better than ±10%. Electroweak Z j j production has been observed, and used to obtain anomalous TGC constraints complementary to those from diboson production. Evidence for electroweak W ± W ± j j production has been obtained, and used to constrain anomalous WWVV QGCs. These latter measurements represent the first steps in studies of VBF and VBS processes, and will be followed by analyses of other single boson and multiboson plus jets channels. √ Looking ahead to Run II, diboson cross sections at s = 13 TeV are expected to increase by a factor of about 2-3, while VBF and VBS cross sections should increase by a factor of about 4-5, improving further the sensitivity to anomalous TGCs and QGCs and providing more precise tests of VBF and VBS processes.
5
References 0
1
2
3
4
5
6
7
8
9 |Δyjj|
Figure 12: The distribution of the rapidity gap |Δy j j | for the jet-jet system in W ± W ± j j events. The vertical dashed line indicates the position of the cut |Δy j j | > 2.4 [22]
The observed number of events in the “VBS” region is used to put the first constraints on possible anomalous contributions to WWVV QGCs. The anomalous QGC contribution is computed using the WHIZARD event generator [23] with a K-matrix unitarisation method [24], and depends on two parameters, α4 and α5 . The first limits on these parameters are obtained; the 95% confidence level allowed regions are −0.14 < α4 < +0.16 (assuming α5 = 0) and −0.23 < α5 < +0.24 (assuming α4 = 0). 7. Conclusions Preliminary W + W − , ZZ and WZ diboson cross section √ measurements are now available from ATLAS at s = 8 TeV, adding to the published 7 TeV cross sections for W + W − , ZZ, WZ, Wγ and Zγ production. These measurements are all in agreement with the corresponding SM predictions, though the 8 TeV W + W − cross section is 2.1σ above expectation. Constraints on anomalous TGCs have been obtained using the 7 TeV data, but not
[1] [2] [3] [4] [5] [6]
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24]
ATLAS Collaboration, JINST 3 S08003 (2008). L. Evans and P. Bryant, JINST 3 S08001 (2008). ATLAS Collaboration, Phys. Rev. D 87 112001 (2013). J. M. Campbell and R. K. Ellis, Phys. Rev. D 62 114012 (2000). ATLAS Collaboration, ATLAS-CONF-2014-033 (2014). T. Melia, P. Nason, R. R¨ontsch and G. Zanderighi, J. High Energy Phys. 11 078 (2011), S. Alioli, P. Nason, C. Oleari and E. Re, J. High Energy Phys. 06 043 (2010), P. Nason, 11 040 (2004). T. Binoth et al., J. High Energy Phys. 12 046 (2006), LHC Higgs Cross Section Working Group Collaboration, arXiv:1307.1347 [hep-ph]. H.-L. Lai et al. , Phys. Rev. D 82 074024 (2010). ATLAS Collaboration, Phys. Rev. Lett. 109 012001 (2012). ATLAS Collaboration, J. High Energy Phys. 03 128 (2013). ATLAS Collaboration, Eur. Phys. Jour. C 72 2173 (2012). ATLAS Collaboration, Phys. Rev. D 87 112003 (2013). ATLAS Collaboration, ATLAS-CONF-2013-020 (2013). ATLAS Collaboration, ATLAS-CONF-2013-021 (2013). see for example: ATLAS Collaboration, Phys. Lett. B 726 88 (2013). ATLAS Collaboration, Phys. Rev. Lett. 112 231806 (2014). A. Belyaev, N. D. Christensen and A. Pukhov, Comput. Phys. Commun. 184 1729 (2013). K. Nakamura (Particle Data Group), J. Phys. G 37 075021 (2010). ATLAS Collaboration, J. High Energy Phys. 04 031 (2014). T. Gleisberg et al., J. High Energy Phys. 02 007 (2009), ATLAS Collaboration, arXiv:1405.6241 (2014). W. Kilian, T. Ohl and J. Reuter, Eur. Phys. J. C 71 1742 (2011), M. Moretti, T. Ohl and J. Reuter, arXiv:0102195 [hep-ph]. A. Alboteanu, W. Kilian and J. Reuter, J. High Energy Phys. 11 010 (2008).
Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246 www.elsevier.com/locate/nppp
Recent Electroweak Results from ATLAS Richard Batleya , on behalf of the ATLAS Collaboration a Department of Physics, University of Cambridge Cavendish Laboratory, J J Thomson Avenue, Cambridge CB3 0HE, UK.
Abstract √ √ Recent electroweak physics results obtained using proton-proton collision data at s = 7 TeV and s = 8 TeV recorded by the ATLAS experiment at the LHC are reviewed. The topics covered include: measurements of diboson cross sections, a measurement of the branching ratio for the rare decay Z → 4, and studies of vector boson fusion and vector boson scattering via measurements of the electroweak components of Z j j and W ± W ± j j production. The constraints on anomalous triple and quartic gauge couplings resulting from these measurements are also summarised. Keywords: LHC, ATLAS, electroweak 1. Introduction Electroweak interactions in the Standard Model (SM) derive from a local SU(2)×U(1) gauge symmetry. The expected triple (TGC) and quartic (QGC) gauge boson couplings can be tested at the LHC through studies of multiboson (W, Z, γ) production or through studies of vector boson fusion (VBF) and vector boson scattering (VBS) processes. Studies of VBS processes are also of interest as they can potentially test whether the single Higgs scalar boson of the SM is indeed solely responsible for unitarising the amplitude for longitudinal W boson scattering, WL WL → WL WL . Electroweak physics results from proton-proton collision data recorded by the ATLAS experiment [1] during √ √ the 2011 ( s = 7 TeV) and 2012 ( s = 8 TeV) runs of the LHC [2] are reviewed below. Emphasis is placed on the most recent results: a preliminary measurement of √ the W + W − cross section at s = 8 TeV; a measurement of the branching ratio for the rare decay Z → 4; and studies of VBF and VBS processes via measurements of electroweak Z j j and W ± W ± j j production. 2. W +W − production At the LHC, W + W − production occurs dominantly via the process qq → W + W − (see Fig. 1) and is sensitive to the WWZ and WWγ TGCs. Additional contributions from gg → W + W − or gg → H → W + W − contribute about 2% and 7%, respectively, to the total W + W − cross section. The W + W − cross section measured using the ATLAS 7 TeV data set [3], 51.9 ± 2.0(stat) ± http://dx.doi.org/10.1016/j.nuclphysbps.2015.01.051 2405-6014/© 2015 Published by Elsevier B.V.
3.9(stat) ± 2.0(lumi) pb, is in agreement with the predicted SM cross section of 44.7 ± 2.1 pb computed at next-to-leading order (NLO) QCD using MCFM [4]. A preliminary√ ATLAS measurement of the W + W − cross section at s = 8 TeV has now also been obtained [5].
q¯
W
q¯
W
q¯
q
Z/γ ∗
W
q
TGC vertex
W
Figure 1: Leading order Feynman diagrams for the process qq → W + W −
Candidate W + W − events in the ATLAS 8 TeV data set are selected by requiring two isolated, high pT , oppositely charged electrons or muons (e+ e− , μ+ μ− or e± μ∓ ) (leading lepton pT > 25 GeV, sub-leading pT > 20 GeV), accompanied by large missing transverse enmiss ergy: ET,rel > 45(15) GeV and pmiss > 45(20) GeV, T where the numbers in brackets correspond to the mixedmiss = ETmiss × sin |Δφ| flavour e± μ∓ channel. Here, ET,rel where Δφ is the azimuthal angle between the ETmiss vector and the closest selected jet or lepton in the event; miss rather than ETmiss reduces the impact of misusing ET,rel calibrations or fluctuations in the measurement of jet or lepton energies. The quantity ETmiss is obtained primarily from the calorimeter energies and reconstructed muons, while pmiss is obtained by summing over reconstructed T tracks. In addition, a requirement |Δφmiss | < 0.3(0.6) is placed on the difference in azimuthal directions of
242
Events
10000
s = 8 TeV, ∫ Ldt = 20.3 fb
600
ATLAS Preliminary
s = 8 TeV, ∫ Ldt = 20.3 fb-1
e± ν μ ν channel
Data Top Wjets stat. unc.
WW (× 1.21) Zjets other diboson stat. ⊕ syst. unc.
500 400 300 200 100 50
ATLAS Preliminary
100
150
200
250
300
Data -1
mll [GeV]
WW MC Top MC
e± ν μ ν channel ±
12000
700
±
the ETmiss and pmiss vectors. Events are rejected if they T contain addition electrons or muons with pT > 7 GeV, additional jets with pT > 25 GeV (“jet veto”), or, for the same-flavour (e+ e− , μ+ μ− ) channels, if the invariant mass of the dilepton pair lies within 15 GeV of the Z boson mass. The jet veto requirement is effective at reducing the irreducible background from top quark production, as shown in Fig. 2. Using the ATLAS 8 TeV data set, 6636 events are selected in the zero-jet bin, mostly (5067 events) in the e± μ∓ channel.
Events / 10 GeV
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
Zjets MC
Figure 3: Lepton pair invariant mass distribution for selected W + W − events in
Wjets MC
8000
the e± μ∓ channel [5]
other diboson MC stat. unc.
6000
3. Diboson production and anomalous TGCs
4000 2000 0
1
2
3
4
5
6
7
Jet multiplicity
Figure 2: Jet multiplicity distribution in W + W − events for the e± μ∓ channel, without imposing the jet veto requirement (which selects only the first bin) [5]
Fig. 3 shows the reconstructed dilepton mass distribution for the mixed-flavour (e± μ∓ ) channel. The expectation for the W + W − signal is obtained from the POWHEG MC [6] for the dominant qq → W + W − contribution, and from the GG2WW MC [7] for the gg → W + W − contribution. The expected background from top quark production and from (W/Z)+jets production is estimated from data, while the background from diboson production (other than W + W − ) is estimated from MC. In Fig. 3, the W + W − signal expectation has been scaled by a factor of 1.21 to obtain agreement with data. √ The W + W − total cross section at s = 8 TeV is +5.0 +2.2 measured to be 71.4+1.2 −1.2 (stat)−4.4 (syst)−2.1 (lumi) pb, and is about 2.1σ above the SM expectation of 58.7+3.0 −2.7 pb determined using MCFM for the qq → W + W − and gg → W + W − processes and using [8] for the gg → H → W + W − process. Various contributions neglected here may serve to increase the predicted cross section, including NNLO QCD corrections and a k-factor (possibly as large as a factor 2-3) for the gg → W + W − process. Also, the MCFM prediction is obtained using CT10 [9] parton distribution functions (PDFs); using alternative PDFs such as ATLAS-epWZ12 [10] may increase the predicted cross section by as much as 2.9 pb.
Cross section measurements have been published using the 7 TeV ATLAS data set for various diboson production channels: ZZ [11], WZ [12], WW [3], Zγ [13] and Wγ [13]. In all cases, the measured total cross sections are in agreement with the corresponding SM prediction computed at NLO QCD. The observed event rates at high lepton, photon or Z boson pT are used to place constraints on possible anomalous TGC contributions within a model-independent effective Lagrangian framework. The resulting constraints on the anomalous γ,Z coupling parameters f4,5 and hγ,Z 3,4 for the neutral ZZZ, γZZ vertices (forbidden in the SM) improve on those obtained by the combined LEP experiments by a large factor (typically 5-10). The sensitivities to the anomalous coupling parameters λγ,Z , Δκγ,Z , gZ1 for the charged vertices WWZ, WWγ have not yet surpassed those from LEP, though this should happen once analysis of the 8 TeV data is complete. At present, at 8 TeV, besides the W + W − cross section above, preliminary cross section measurements are available for ZZ production in the + − + − ( = e, μ) channel [14], and for WZ production [15]. 4. The rare decay Z → 4 The 4 mass spectrum obtained in observations of the Higgs boson resonance H → ZZ ∗ → 4 [16] shows a second, nearby resonant peak at the Z mass due to Z → 4 decay. In the SM, the Z → 4 resonance arises through s-channel qq → 4 diagrams such as that in Fig. 4(a), where a lepton from the Z decay radiates a virtual Z boson or photon which undergoes internal conversion
243
to a charged lepton pair. The contribution to 4 production in the Z resonance region from t-channel diagrams such as Fig. 4(b) is small, about (3-4)%, while the process gg → 4 contributes only about 0.1%. Unlike the decay H → 4, which involves only a virtual Z boson, the virtual photon component of the dominant s-channel Z → 4 process leads to a final state dominated by the presence of a low mass e+ e− or μ+ μ− pair. To measure the branching ratio BR(Z → 4) as inclusively as possible and with greater statistical precision, a looser cut, m > 5 GeV, than in the Higgs analysis is placed on the invariant mass of all same-flavour opposite-charge lepton pairs (e+ e− , μ+ μ− ) in the 4 system [17]. The minimum pT requirement on electrons and muons is lowered to 7 GeV and 4 GeV, respectively. q
−
120 ATLAS s = 7 TeV, 4.5 fb-1
100
s = 8 TeV, 20.3 fb-1
80
Data Z 10xBkg σstat+syst Z→4l
60 40 20 0 75
80
85
90
95
100
105
m4l [GeV] Figure 5: The invariant mass of the 4 system for Z → 4 candidate events, without imposing the requirement 80 < m4 < 100 GeV [17]
− Z ∗ /γ ∗
q
Events / 3 GeV
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
+ Z/γ ∗
Z ∗ /γ ∗ q¯
+
−
Z ∗ /γ ∗
−
+
q¯
+
Figure 4: Examples of (a) s-channel and (b) t-channel leading order Feynman diagrams contributing to qq → 4 production
Candidate Z → 4 decays are selected in the fourlepton invariant mass range 80 < m4 < 100 GeV, in the e+ e− e+ e− , μ+ μ− μ+ μ− and e+ e− μ+ μ− channels. A total of 152 events is selected for the 7 TeV and 8 TeV data sets and for all channels combined. The m4 distribution for the selected events is shown in Fig. 5. The total background from Z+jets and top quark production, estimated using data-driven methods, and from diboson production, estimated using MC samples, is about 1.0% of selected events. Interference between the s-channel and t-channel contributions is studied at leading order using CALCHEP [18], and found to be small, about 0.2%. The resonant s-channel contribution can therefore effectively be isolated by subtracting the expected fraction of events, fnr , due to the non-resonant t-channel and gg → 4 processes. The Z → 4 branching ratio is then determined by normalising to the observed rate of Z → μ+ μ− events in the same data sets: N 4 2μ , N 2μ 4 where N 4 and N 2μ are the number of backgroundsubtracted Z → 4 and Z → μ+ μ− events observed in the data, 4 and 2μ are the overall selection efficiencies for Z → 4 and Z → μ+ μ− events, respectively, and BR(Z → μ+ μ− ) = (3.366 ± 0.007)% [19]. Normalising to the Z → μ+ μ− data cancels the luminosity BR(Z → 4) = BR(Z → μ+ μ− )(1 − fnr )
uncertainty, and partially cancels the theoretical uncertainties. The branching ratio is determined for the phase space region defined by the requirements 80 < m4 < 100 GeV and m > 5 GeV for all e+ e− and μ+ μ− pairs in the 4 system. In this region, the Z → 4 branching ratio is measured to be (3.20 ± 0.25 ± 0.13) × 10−6 , in agreement with the SM prediction from POWHEG of (3.33 ± 0.01) × 10−6 . Interference between the schannel and t-channel contributions (0.2%) is included in the systematic error on the measured branching ratio. 5. Vector boson fusion: Z j j production The VBF production, qq → qqZ, of a Z boson via the fusion of two W bosons, illustrated in Fig. 6(a), gives rise to two final state jets from the scattered quarks which tend to be produced in the forward directions, at large rapidities. The VBF process is accompanied by other electroweak Z j j diagrams, and is dominated by strong (QCD) Z j j production processes such as the diagram shown in Fig. 6(b). The two jets in the Z j j final state tend to be produced with higher pT in the electroweak case, and tend to have larger invariant mass, m j j , and larger rapidity separation, |Δy j j |. Furthermore, the lack of any colour exchange between the scattered quarks in the electroweak VBF process results, on average, in less additional jet activity in the region between the two forward jets than for the QCD contributions. The selection of candidate Z j j events [20] requires at least two jets, with pT > 55(45) GeV for the leading (sub-leading) jet. The Z boson is reconstructed via the decays Z → e+ e− and Z → μ+ μ− , where the decay leptons are required to have pT > 25 GeV and in-
244
q
−
g
−
μ ,e
W+
jet
q
q
−
−
μ ,e
Z
Z
W−
+
+
μ ,e
1 dσ σ dNgap
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
+
0.6
Data (2012) Sherpa Zjj (QCD + EW)
0.5
Sherpa Zjj (QCD) Powheg Zjj (QCD + EW)
+
μ ,e
Powheg Zjj (QCD)
0.4 q
q
q
g
0.3
ATLAS
∫ L dt = 20.3 fb
-1
Figure 6: Examples of (a) electroweak VBF and (b) QCD Feynman diagrams
s = 8 TeV high-mass region
0.2
contributing to Z j j production
106
ATLAS
baseline region -1 ∫ L dt = 20.3 fb s = 8 TeV
105 104
Data (2012) QCD Zjj EW Zjj VZ Top Multijets
1
1.5 1 0.5 0
1
2
3
4
5
6
7
8
gap
Njet
103
gap Figure 8: The unfolded Njet distribution for the Z j j “high mass” region. The
102
expected distribution neglecting the electroweak Z j j component is given by the horizontal dashed lines [20]
10 Prediction Data
1.5
0.5
Powheg Data
Nobs/200 GeV
variant mass m in the range 81 < m < 101 GeV. The distribution of the invariant mass m j j is plotted in Fig. 7, and shows the enhanced electroweak component at high m j j , but also that the QCD component remains dominant even at high mass.
Sherpa Data
0.1
2 1.5 1 0
500
1000
1500
2000
2500 3000 mjj [GeV]
Figure 7: The m j j distribution in Z j j candidate events, before imposing any gap
requirements on m j j or Njet [20]
The “search” region is defined by the additional regap gap quirements m j j > 250 GeV and Njet = 0, where Njet is the number of additional reconstructed jets with pT > 25 GeV in the rapidity interval between the two leading jets. The “high mass” region is defined by increasing the jet-jet mass cut to m j j > 1 TeV and removing the gap Njet requirement. The electroweak component is expected to be about 4.0% (12%) of selected events for the “search” (“high mass”) regions, while backgrounds from diboson and top quark production are estimated to be small, about 1.3% (3%). Background-subtracted, unfolded differential distributions are obtained for a vagap riety of kinematic variables; the unfolded Njet distribution for events in the “high mass” region is shown in gap Fig. 8, for example. The Njet distribution is reasonably well modelled by both the (LO) SHERPA [21] and (NLO) POWHEG MCs. The reduced jet activity expected for the electroweak component is visible in Fig. 8 from a comparison of the solid and dashed distributions.
To extract the Z j j electroweak component (Z j j − EW), a template fit is made to the m j j distribution for selected events in the “search” region. The template for the Z j j − EW signal is obtained from the SHERPA MC, while the background template is obtained from the SHERPA MC for the QCD Z j j process, plus small diboson and top quark contributions from other MC samples. To account for possible mismodelling, the background template is constrained using a data “control” region obgap tained by reversing the jet veto (Njet ≥ 1). The fit to the m j j distribution is shown in Fig. 9, and gives a Z j j−EW cross section of 54.7 ± 4.6(stat)+9.8 −10.4 (syst) ± 1.5(lumi) fb, in agreement with the SM expectation computed using POWHEG. The background-only hypothesis is excluded at over 5σ, and the measurement thus represents an observation of the electroweak Z j j process. The subset of “search” region events with m j j > 1 TeV (900 events, with an expected Z j j − EW contribution of about 260 events) is used to constrain anomalous contributions to the WWZ triple gauge vertex. Limits are placed on the anomalous coupling parameters gZ1 and λZ which are less restrictive than those obtained to date from WZ production, but which are complementary in that the two W bosons involved in the Z j j VBF process are spacelike, not timelike.
245
10
103
ATLAS L dt = 20.3 fb-1 s = 8 TeV search region
102 10 Data (2012) Background
BKG BKG + EW Data Data
1
Background + EW Zjj
1.5 1 0.5 1 0.5 0
constrained unconstrained
500
1000 1500 2000 2500 3000 3500 mjj [GeV]
Figure 9: Results of the fit to the m j j distribution in Z j j events used to extract the electroweak Z j j component. The horizontal dashed lines in the lower panel show the effect of removing the “control” region constraint on the background template [20]
6. Vector boson scattering: W ±W ± j j production Examples of electroweak VBS contributions to the process qq → WW j j, involving WWVV (V = W, Z) QGCs or Higgs exchange, are shown in Fig. 10. Similarly to the case of Z j j production above, the VBS processes are accompanied by other electroweak contributions and are dominated by strong (QCD) contributions. Again, the VBS diagrams for WW j j production tend to produce two high pT , forward jets, and the VBS contribution can be enhanced by requiring large invariant mass, m j j , or large rapidity separation, |Δy j j |, of the two forward jets. Unlike Z j j production, and since at LO the two W bosons produced in gg → WW j j processes must have opposite electric charge, the QCD contribution to WW j j production can be reduced by a large factor by requiring that the two W bosons have the same charge.
q
q
q
q
W+
W+
q
q
q
102
10
ATLAS 20.3 fb-1, s = 8 TeV
Data 2012 Syst. Uncertainty W±W±jj Electroweak W±W±jj Strong Prompt Conversions Other non-prompt
1
10-1
5
Data/Bkg Bkg Uncertainty (Sig+Bkg)/Bkg
0 200 400 600 800 1000 1200 1400 1600 1800 2000 mjj [GeV]
Figure 11: The distribution of the invariant mass m j j of the jet-jet system in W ± W ± j j events. The vertical dashed line indicates the position of the cut m j j > 500 GeV [22]
W+
q
data [22] by requiring two jets with pT > 30 GeV, with jet-jet invariant mass m j j > 500 GeV. In addition, events are required to contain exactly two same-sign electrons or muons (e± e± , μ± μ± or e± μ± ), each with pT > 25 GeV, and to contain large missing transverse energy ETmiss > 40 GeV. To reduce backgrounds from diboson and top quark production, events with additional leptons (electrons with pT > 7 GeV or muons with pT > 6 GeV), or with reconstructed b-quark jets (with pT > 30 GeV and |η| < 2.5) are rejected. To reduce the background from (Z → e+ e− )+jets production, where the charge of one of the electrons is misreconstructed, events with e± e± mass within 10 GeV of the Z boson mass are removed. The m j j distribution for W ± W ± j j candidates for all channels combined is shown in Fig. 11. The signal expectation is evaluated using events generated with SHERPA and normalised to the cross section computed using POWHEG. Electroweak W ± W ± j j production becomes the dominant expected contribution at high values of m j j . The m j j > 500 GeV requirement leaves a sample of 50 events in the 8 TeV data, with an estimated irreducible background of about 20 events. The background-only hypothesis is excluded at 4.5σ, compared with an expected exclusion significance of 3.4σ.
Events/50 GeV
∫
4
Data/Background
Nobs / 250 GeV
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
W+
Figure 10: Examples of electroweak VBS Feynman diagrams contributing to W ± W ± j j production
Same-sign W ± W ± j j events are selected in the 8 TeV
Fig. 12 shows the |Δy j j | distribution for the sample of 50 selected W ± W ± j j candidates. The electroweak WW j j component is extracted using the “VBS” region, defined by imposing the additional requirement |Δy j j | > 2.4, leaving a data sample of 34 events. Af-
246
R. Batley / Nuclear and Particle Physics Proceedings 258–259 (2015) 241–246
Events
ter subtracting the estimated irreducible background of about 15 events, and subtracting also the strong WW j j contribution of about 1.3 events, the electroweak WW j j cross section (including interference between the electroweak and strong WW j j contributions) is measured to be 1.3 ± 0.4(stat) ± 0.2(syst) fb, in good agreement with the SM expectation of 0.95 ± 0.06 fb computed using POWHEG. The background-only hypothesis is excluded with a significance of 3.6σ, compared with an expected significance of 2.8σ, and as such the measurement provides the first evidence for electroweak WW j j production. 30 25
ATLAS 20.3 fb-1, s = 8 TeV mjj > 500 GeV
Data 2012 Syst. Uncertainty W±W±jj Electroweak W±W±jj Strong Prompt Conversions Other non-prompt
20 15 10
yet the 8 TeV data. Even so, experimental constraints on anomalous ZZZ and ZZγ couplings are now dominated by those obtained at the LHC, while the constraints obtained on anomalous WWZ and WWγ couplings are already approaching the combined constraints from the LEP experiments. The branching ratio for the rare decay Z → 4 has been determined with a precision better than ±10%. Electroweak Z j j production has been observed, and used to obtain anomalous TGC constraints complementary to those from diboson production. Evidence for electroweak W ± W ± j j production has been obtained, and used to constrain anomalous WWVV QGCs. These latter measurements represent the first steps in studies of VBF and VBS processes, and will be followed by analyses of other single boson and multiboson plus jets channels. √ Looking ahead to Run II, diboson cross sections at s = 13 TeV are expected to increase by a factor of about 2-3, while VBF and VBS cross sections should increase by a factor of about 4-5, improving further the sensitivity to anomalous TGCs and QGCs and providing more precise tests of VBF and VBS processes.
5
References 0
1
2
3
4
5
6
7
8
9 |Δyjj|
Figure 12: The distribution of the rapidity gap |Δy j j | for the jet-jet system in W ± W ± j j events. The vertical dashed line indicates the position of the cut |Δy j j | > 2.4 [22]
The observed number of events in the “VBS” region is used to put the first constraints on possible anomalous contributions to WWVV QGCs. The anomalous QGC contribution is computed using the WHIZARD event generator [23] with a K-matrix unitarisation method [24], and depends on two parameters, α4 and α5 . The first limits on these parameters are obtained; the 95% confidence level allowed regions are −0.14 < α4 < +0.16 (assuming α5 = 0) and −0.23 < α5 < +0.24 (assuming α4 = 0). 7. Conclusions Preliminary W + W − , ZZ and WZ diboson cross section √ measurements are now available from ATLAS at s = 8 TeV, adding to the published 7 TeV cross sections for W + W − , ZZ, WZ, Wγ and Zγ production. These measurements are all in agreement with the corresponding SM predictions, though the 8 TeV W + W − cross section is 2.1σ above expectation. Constraints on anomalous TGCs have been obtained using the 7 TeV data, but not
[1] [2] [3] [4] [5] [6]
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24]
ATLAS Collaboration, JINST 3 S08003 (2008). L. Evans and P. Bryant, JINST 3 S08001 (2008). ATLAS Collaboration, Phys. Rev. D 87 112001 (2013). J. M. Campbell and R. K. Ellis, Phys. Rev. D 62 114012 (2000). ATLAS Collaboration, ATLAS-CONF-2014-033 (2014). T. Melia, P. Nason, R. R¨ontsch and G. Zanderighi, J. High Energy Phys. 11 078 (2011), S. Alioli, P. Nason, C. Oleari and E. Re, J. High Energy Phys. 06 043 (2010), P. Nason, 11 040 (2004). T. Binoth et al., J. High Energy Phys. 12 046 (2006), LHC Higgs Cross Section Working Group Collaboration, arXiv:1307.1347 [hep-ph]. H.-L. Lai et al. , Phys. Rev. D 82 074024 (2010). ATLAS Collaboration, Phys. Rev. Lett. 109 012001 (2012). ATLAS Collaboration, J. High Energy Phys. 03 128 (2013). ATLAS Collaboration, Eur. Phys. Jour. C 72 2173 (2012). ATLAS Collaboration, Phys. Rev. D 87 112003 (2013). ATLAS Collaboration, ATLAS-CONF-2013-020 (2013). ATLAS Collaboration, ATLAS-CONF-2013-021 (2013). see for example: ATLAS Collaboration, Phys. Lett. B 726 88 (2013). ATLAS Collaboration, Phys. Rev. Lett. 112 231806 (2014). A. Belyaev, N. D. Christensen and A. Pukhov, Comput. Phys. Commun. 184 1729 (2013). K. Nakamura (Particle Data Group), J. Phys. G 37 075021 (2010). ATLAS Collaboration, J. High Energy Phys. 04 031 (2014). T. Gleisberg et al., J. High Energy Phys. 02 007 (2009), ATLAS Collaboration, arXiv:1405.6241 (2014). W. Kilian, T. Ohl and J. Reuter, Eur. Phys. J. C 71 1742 (2011), M. Moretti, T. Ohl and J. Reuter, arXiv:0102195 [hep-ph]. A. Alboteanu, W. Kilian and J. Reuter, J. High Energy Phys. 11 010 (2008).