CERN-PH-EP-2011-122Measurement of the W\rightarrow\tau\nu_{\tau} Cross Section in pp Collisions at \sqrt{s} = 7 TeV with the ATLAS experiment

Cern-Ph-Ep-2011-122 Measurement of the Cross Section in Collisions at = 7 TeV with the ATLAS experiment

The ATLAS Collaboration
Abstract

The cross section for the production of bosons with subsequent decay is measured with the ATLAS detector at the LHC. The analysis is based on a data sample that was recorded in 2010 at a proton-proton center-of-mass energy of and corresponds to an integrated luminosity of 34 pb. The cross section is measured in a region of high detector acceptance and then extrapolated to the full phase space. The product of the total production cross section and the branching ratio is measured to be .

keywords:
W boson, Standard Model, tau lepton
journal: Physics Letters B

1 Introduction

The study of processes with leptons in the final state is an important part of the ATLAS physics program, for example in view of searches for the Higgs boson or supersymmetry AtlasCSCBook (); paper:AHtautau (); note:tHplus (). Decays of Standard Model particles to leptons, in particular and , are important background processes in such searches. Studies of the decay complement the measurement of W production in the muon and electron decay modes ATLAS-CONF-2011-041 (); WZnewpaper (). In addition, decays can be used to validate the reconstruction and identification techniques for leptons and the measurement of the missing transverse energy (), which are both fundamental signatures in a wide spectrum of measurements at the LHC.

At next-to-next-to-leading order (NNLO), the signal is predicted to be produced at with a cross section times branching ratio of  ref:FEZW (); ref:MSTW2008 (); Aad:2010yt (). Since purely leptonic decays cannot be easily distinguished from electrons and muons from or decays, the analysis presented in this paper uses only hadronically decaying leptons (). Events from production contain predominantly low- bosons decaying into leptons with typical visible transverse momenta between 10 and 40 GeV. In addition, the distribution of the missing transverse energy, associated with the neutrinos from the and decays, has a maximum around 20 GeV and a significant tail up to about 80 GeV.

Previous measurements at hadron colliders of boson production with the subsequent decay based on collisions were reported by the UA1 collaboration ref:WtaunuUA1 () at center-of-mass energies of = 546 GeV and = 630 GeV and by the CDF and D0 collaborations ref:WtaunuCDF (); ref:WtaunuD0 () at a center-of-mass energy of = 1.8 TeV.

In this Letter, we describe the measurement of this process with  TeV collision data, which were recorded with the ATLAS experiment at the LHC.

2 The ATLAS detector

The ATLAS detector is described in Ref. ATLASDetector (). The cylindrical coordinate system is defined with polar angles  relative to the beamline and azimuthal angles  in the plane transverse to the beam. Pseudorapidities are defined as . Transverse momenta, , are defined as the component of momentum perpedicular to the beamline. Distances are measured in the - plane as .

Measurements of charged-particle trajectories and momenta are performed with silicon detectors in the pseudorapidity range , and also by a straw-tube tracking chamber in the range . Together, these systems form the inner tracking detector, which is contained in a 2 T magnetic field produced by a superconducting solenoid. These tracking detectors are surrounded by a finely segmented calorimeter system which provides three-dimensional reconstruction of particle showers up to . The electromagnetic calorimeter uses liquid argon as the active material and comprises separate barrel (), end-cap () and forward () components. The hadron calorimeter is based on scintillating tiles in the central region (). It is extended up to by end-caps and forward calorimeters which use liquid argon. The muon spectrometer measures the deflection of muon tracks in the field of three large superconducting toroidal magnets. It is instrumented with trigger and high-precision tracking chambers.

The trigger system consists of three levels. The first level is implemented as a hardware trigger, while the decision on the following levels is based on software event processing similar to the offline reconstruction.

3 Data samples

The data used in this measurement were recorded in proton-proton collisions at a center-of-mass energy of  TeV during the 2010 LHC run. The integrated luminosity of the data sample, considering only data-taking periods where all relevant detector subsystems were fully operational, is 34 pb lumipaper (); lumi (). The data were collected using triggers combining the two main signatures of decays, namely the presence of a hadronically decaying lepton and missing transverse energy.

Processes producing or bosons that subsequently decay into electrons or muons constitute important backgrounds to this measurement if the lepton from the decay or an accompanying jet is misidentified as a hadronically decaying lepton. Here, the missing transverse energy signature arises from a decay neutrino or the misreconstruction of jets or of other objects in the event. Also, decays with the decaying leptonically are considered as a background. Incompletely reconstructed and decays can also enter the signal sample. The number of background events from these electroweak processes is referred to as in the following.

The production of and bosons in association with jets is simulated with the PYTHIA Pythia () generator with the modified LO parton distribution function (PDF) MRSTLO* ref:mrstlo () and normalized to the NNLO cross section; processes are generated with MC@NLO MCatNLO (), where parton showers and hadronization are simulated with HERWIG Herwig () and the underlying event with JIMMY JIMMY (). The TAUOLA TAUOLA () and PHOTOS PHOTOS () programs are used to model the decay of leptons and the QED radiation of photons, respectively.

All simulated samples include multiple proton-proton interactions (pile-up) produced with PYTHIA using the ATLAS MC10 tune pileup1 (). Those samples are passed through a full detector simulation based on GEANT4 Geant4 (); ATLASSIM (). The simulated events are re-weighted so that the distribution of the number of reconstructed primary vertices per bunch crossing matches the data.

Due to their large production cross sections, QCD processes provide a significant background if quark/gluon jets (QCD jets) are misidentified as hadronic decays and a significant amount of is measured, mainly due to incomplete reconstruction. The number of QCD background events is estimated directly from data.

4 Object reconstruction

Electron candidates, which together with muons are relevant for the electroweak background, are reconstructed from a cluster in the electromagnetic calorimeter matched to a track in the inner tracking detector. The cluster must have a shower profile consistent with an electromagnetic shower ATL-PHYS-PUB-2011-006 (). Muon candidates are reconstructed by combining tracks in the muon spectrometer with tracks in the inner tracking detector ATLAS-CONF-2011-063 ().

Jets are reconstructed with the anti- algorithm AntiKT () with a radius parameter . The jet energies are calibrated ATLASJETEnergyScale () using a - and -dependent calibration scheme, corrected for losses in dead material and outside the jet cone LocalHadronCalib (). All jets considered in this analysis are required to have a transverse momentum above  GeV and a pseudorapidity in the range .

Reconstructed jets within provide the starting point (seed) for the reconstruction of hadronic decays. The direction of a  candidate is taken directly from the corresponding seed jet. The energy is calibrated by applying a dedicated correction extracted from Monte Carlo to the sum of energies of the cells that form the clusters of the seed jet ATLAS-CONF-2011-077 (). Therefore, the energy of the refers to the visible decay products. The transverse momentum is calculated as , i.e.  candidates are treated as massless. Good-quality tracks are associated with a  candidate if they are found within around the seed jet axis. At least one track must be associated to the candidate.

The identification ATLAS-CONF-2011-077 () is based on eight observables: The invariant mass of the decay products is calculated separately using the associated tracks and the associated clusters. The fact that the decay products are typically more collimated than QCD jets is quantified by calculating the transverse momentum-weighted radius from tracks and the energy-weighted radius from electromagnetic energy information. The fraction of transverse energy within 0.1 of the seed direction is used as well. Further discrimination is provided by the fraction of the transverse momentum carried by the highest- track and the fraction of transverse energy deposited in the electromagnetic calorimeter, for which higher values are expected in case of hadronic decays compared to QCD jets. For candidates with more than one associated track, the lifetime is also exploited by measuring the decay length significance of the associated secondary vertex in the transverse plane. The single most discriminating of these quantities is the energy-weighted radius

(1)

where iterates over cells in the first three layers of the electromagnetic calorimeter associated with the candidate, is defined relative to the seed axis, and is the cell transverse energy.

These eight variables are combined in a boosted decision tree discriminator (BDT) boosting (), which provides an output value between 0 (background-like) and 1 (signal-like) with a continuous gradient of signal and background efficiency. This discriminator was optimized using a combination of and Monte Carlo samples for the signal. The background was modeled from dijet events selected from data. For  transverse momenta () above 20 GeV, the efficiency of the identification at the tighter working point of the BDT identification considered for this measurement is about 30% with a jet rejection factor of 100 for candidates with one track, while for candidates with three tracks it is about 35% with a rejection factor of 300 ATLAS-CONF-2011-077 (). Additional requirements on the calorimeter and tracking properties of candidates are used to discriminate against electrons and muons.

The missing transverse energy in the event, , is reconstructed as , where is the vector sum of all calorimeter energy clusters in the region , corrected for identified muons ref:perf (). With good approximation, the resolution of components is proportional to , where the scaling factor depends on both the detector and reconstruction performance and is calculated from all calorimeter energy clusters. The factor is about 0.5 for minimum bias events ATLAS-CONF-MET-Calibration-7TeV ().

In order to reject events with large reconstructed due to fluctuations in the energy measurement, we define the significance of as:

(2)

is found to provide better discrimination between the signal and the background from QCD jets than a simple requirement.

A B C D
(Data) 2335 4796 1577 27636
() 1811 25 683 16 269 8 93 5
284 7 118 4 388 9 90 4
0.38 0.01 0.31 0.01 0.087 0.003
127 8 3953 75 885 45 27444 166
Table 1: Estimated sample compositions and factors (as defined in Equation 4) in the signal region A and control regions B, C, and D defined in the text.

5 Event selection

Events are selected using triggers based on the presence of a jet and . In the earlier part of the 2010 data taking, corresponding to an integrated luminosity of 11 pb, a loosely identified candidate with GeV (as reconstructed at the trigger level) in combination with GeV was required. In the second part of the period (24 pb), a tighter identification and higher thresholds of 16 GeV and 22 GeV had to be used for and , respectively, due to the increasing luminosity. The signal efficiencies of these two triggers with respect to the offline selection are estimated from the simulation to be (81.30.8)% and (62.70.7)%, respectively.

Events satisfying the trigger selection are required to have at least one reconstructed vertex that is formed by three or more tracks with 150 MeV. Further selection requirements based on calorimeter information are applied to reject non-collision events and events containing jets that were incompletely reconstructed or significantly affected by electronic noise in the calorimeters.

The calorimeter has a lower resolution for jets in the barrel-endcap transition regions. In order to ensure a uniform resolution, events are rejected if a jet or a candidate with is found. In events where the is found to be collinear to one of the jets, the reconstructed is likely to originate from an incomplete reconstruction of this jet. Therefore, a minimum separation rad is required.

In order to suppress backgrounds from other leptonic and decays, events containing identified electrons or muons with are rejected. The highest- identified candidate in the event is considered for further analysis and required to be in the pseudorapidity range and to have GeV. A minimum of 30 GeV is required and events are rejected if .

6 Background estimation

The number of expected events from signal and electroweak background processes is obtained from simulation. This is justified by the good agreement between data and simulation observed in the ATLAS cross section measurements ATLAS-CONF-2011-041 (); WZnewpaper () through decays into electrons or muons. It is further validated using a high-purity data sample of events, in which the muon is removed and replaced by a simulated lepton. Thus, only the decay and the corresponding detector response are taken from simulation while the underlying kinematics and all the other properties of the event are obtained from the events selected in data. Figure 1 compares the distribution of for the -embedded data sample with simulated events. A good agreement is observed within the statistical uncertainties, which adds further confidence in the electroweak background event model provided by the simulated event samples used in this analysis.

Figure 1: Distribution of for the -embedded data sample (points) and simulated events (histogram), including statistical uncertainties.

The background contribution from QCD jet production, for which the cross section is large and the selection efficiency is low, cannot be reliably modeled using simulated events alone and is thus estimated from data. In addition to the signal-dominated data set defined by the selection described in Section 5, three background control regions are defined by inverting the requirements on the and/or the identification (ID), resulting in the following four samples:

  • Region A: and candidates satisfying the signal ID requirements described in Section 4;

  • Region B: and candidates satisfying the signal-region ID requirements;

  • Region C: and candidates satisfying a looser -ID but failing the signal-region ID requirements;

  • Region D: and candidates satisfying a looser -ID but failing the signal-region ID requirements.

Here, the looser -ID region is defined by selecting candidates with a lower value of the BDT output.

After ensuring that the shape of the distribution for the QCD background is independent of the -ID requirement and assuming that the signal and electroweak background contributions in the three control regions are negligible, an estimate for the number of QCD background events in the signal region A is provided by

(3)

where represents the number of observed events in region .

In order to take into account the residual signal and EW background contamination in the control regions, the number of selected events, , needs to be replaced in Equation 3 by , where

(4)

is the ratio of simulated signal and EW background events in the control region and the signal region. Therefore Equation 3 becomes:

The statistical error on includes both the uncertainty on the calculation of the coefficients, due to the Monte Carlo statistics, and the statistical uncertainty of the data in the four regions. The resulting estimates of the sample compositions are summarized in Table 1.

Figure 2: (a) distribution in the combined region AB, extended over the full range. The QCD background shape has been extracted from regions CD. Monte Carlo signal and EW background in regions AB are also shown; (b) the identification variable in the combined region AC. The QCD background shape has been extracted from regions BD. Monte Carlo signal and EW background in regions AC are also shown.
Figure 3: (a) Distribution of missing transverse energy in signal region A on a linear scale. The QCD background shape has been extracted from control region C. (b) Same distribution on a logarithmic scale. (c) Transverse momentum and (d) number of tracks of candidates in signal region A. The QCD background shape has been extracted from control region B. (e) Distribution of and (f) transverse mass m in signal region A. The QCD background shape has been extracted from control region C. The expectation from Monte Carlo signal and EW background in region A are also shown.

The quality of the description of the selected data by the background models can be judged from Figures 2 and 3, where data and the background estimates (EW and QCD) are shown. Figure 2 shows the distribution of in regions A and B, extended over the full range, for all events passing the selection criteria except for the requirement. In Figure 2 the distribution of is shown. In this case events passing the selection criteria but considering candidates identified by the loose and the tight selections in regions A and C are shown. The agreement between data and Monte Carlo expectation confirms the results obtained by the data-driven background estimation. In Figure 3 the distribution of , the spectrum, the number of tracks associated to the candidate, the distribution of and the transverse mass, , in the selected signal region A are shown, illustrating the characteristic properties of decays. In all the distributions reasonable agreement is observed between the data and Monte Carlo prediction.

7 Cross section measurement

The fiducial cross section is measured in a phase space region given by the geometrical acceptance of the detector and by the kinematic selection of the analysis (as described in Section 5). This region is defined based on the decay products from a simulated hadronic decay and corresponds to the criteria presented in Table 2.

, excluding
Table 2: Definition of the acceptance region.

Here, the visible momentum and pseudorapidity are calculated from the sum of the four-vectors of the decay products from the simulated hadronic decay, except for the neutrinos. This momentum also includes photons radiated both from the lepton and from the decay products themselves, considering only photons within with respect to the . The minimum requirement translates into a cut on the transverse component of the sum of the simulated neutrino four-vectors .

The fiducial cross section, including the branching ratio , is computed as

(5)

where is the number of observed events in data, is the number of estimated (QCD and EW) background events (signal region A in Table 1), and is the integrated luminosity. is the correction factor that takes into account the efficiency of trigger, reconstruction and identification and the efficiency of all selection cuts within the acceptance:

(6)

where is the number of fully simulated signal events passing the reconstruction, trigger and the selection cuts of the analysis and is the number of simulated signal events within the fiducial region defined above.

With the kinematic and geometrical signal acceptance

(7)

where is the total number of simulated signal events while is the denominator of , the total cross section

(8)

can be obtained. and are determined using a PYTHIA Monte Carlo signal sample described in Section 3. The fiducial acceptance is found to be and the correction factor .

The measured fiducial cross section of the decay is and the total cross section is found to be .

Several alternative analyses are performed to confirm these results. For example, the BDT ID is replaced by a simpler identification based on cuts on three of the ID variables only ATLAS-CONF-2011-077 (). Also, in order to study the influence of pile-up on the result, the signal selection is restricted to events with only one reconstructed primary vertex. In both cases consistent results are found.

8 Systematic uncertainties

Table 3 summarizes the systematic uncertainties. The main sources are discussed in the following.

Trigger efficiency 6.1% 6.1% - 7.0%
Energy scale 6.7% 8.7% - 8.0%
ID efficiency 9.6% 4.1% - 10.3%
Jet misidentification - 7.2% - 1.1%
Electron misidentification - 4.5% - 0.7%
Pile-up reweighting 1.4% 1.2% - 1.6%
Electron reconstruction/identification - 1.2% - 0.2%
Muon reconstruction - 0.3% - 0.04%
Underlying event modeling 1.3% 1.1% - 1.5%
Cross section - 4.5% - 0.7%
QCD estimation: Stability/correlation - - 2.7% 0.2%
QCD estimation: Sig./EW contamination - - 2.1% 0.1%
Monte Carlo statistics 1.4% 2.4% 6.0% 1.5%
Total systematic uncertainty 13.4% 15.2% 6.9% 15.1%
Table 3: Summary table for systematic uncertainties. For the systematic uncertainty on the fiducial cross section measurement, correlations between the systematics affecting and have been taken into account.

Monte Carlo predictions


The trigger efficiency is determined in Monte Carlo for the combined and triggers used in the two data periods. The differences between the measured trigger responses of the two trigger components in data and Monte Carlo are used to determine the systematic uncertainty. A pure and unbiased sample enriched with events is obtained in data by applying an independent () trigger and selected cuts of the event selection like the BDT ID. The corresponding () trigger part is applied to this sample and the response of this trigger is compared to the response in Monte Carlo. The observed differences are integrated over the offline and range used for the cross section measurement. The total systematic uncertainty after the combination of the different trigger parts is 6.1%.

The signal and background acceptance depends on the energy scale of the clusters used in the computation of and and the energy scale of the calibrated candidates. Based on the current knowledge of the calibration the uncertainty due to cluster energy within the detector region is at most 10% for of 500 MeV and within 3% at high  ATLAS-CONF-2011-080 (). In the forward region it is estimated to be 10%. The effect on and has been evaluated by scaling all clusters in the event according to these uncertainties and recalculating and . At the same time, the energy scale has been varied according to its uncertainty ATLAS-CONF-2011-077 (). This uncertainty depends on the number of tracks associated to the candidate, its and the region in which it was reconstructed, and ranges from 2.5% to 10%. In addition, the sensitivity of the signal and background efficiency to the resolution has been investigated ATLAS-CONF-MET-Calibration-7TeV (). Consequently, the yield of signal and EW background varies within 6.7% and 8.7%, respectively.

The identification and reconstruction efficiency of candidates was studied with Monte Carlo and samples and was found to vary with different simulation conditions such as different underlying event models, detector geometry, hadronic shower modeling and noise thresholds for calorimeter cells in the cluster reconstruction. In Ref. ATLAS-CONF-2011-077 (), these uncertainties are evaluated as a function of , separately for candidates with one or multiple tracks and low or high multiplicity of primary vertices in the event. The corresponding changes in the signal and EW background efficiencies are found to be 9.6% and 4.1%, respectively.

The probability of a jet or electron to be misidentified as a candidate has been evaluated in data and compared with the expectation from Monte Carlo. The rate of jets that are misidentified as candidates was calculated using a selection of Wjets events (with ) and measuring the fraction of reconstructed candidates that are found by the identification. The difference of this misidentification rate in Monte Carlo compared to that in data is 30% and this was applied as a systematic uncertainty to the fraction of events mimicked by a jet. The overall uncertainty on the EW background is 7.2%. The misidentification probability of electrons as candidates has been determined with a “tag-and-probe” method using events where the identification and electron veto is applied to one of the electrons. The difference between the misidentification probability in data and Monte Carlo as a function of  has been applied as a systematic uncertainty to candidates mimicked by an electron. It amounts to 4.5% for the total EW background.

Other sources of systematic uncertainty have been evaluated and were found to have only small effects on the resulting cross section measurement, for example the procedure to include pile-up effects, the uncertainty on the lepton selection efficiency entering via the veto of electrons and muons and the influence of the underlying event modeling on quantities. The uncertainties on the cross sections used for the EW background are taken from ATLAS measurements, when available, or theoretical NNLO calculations, and lie between 3 and 9.7% Aad:2010yt (); ref:xsec:top (); ref:FEZW (); ref:MSTW2008 (). The uncertainty on the integrated luminosity is 3.4% lumipaper (); lumi ().

QCD background estimation


Two different sources of systematic uncertainty arising from the method of estimating the QCD background events from data have been studied. The stability of the method and the small correlation of the two variables ( ID and ) used to define the control regions have been tested by varying the threshold. The systematic uncertainty due to the correction for signal and EW background contamination in the control regions was obtained by varying the fraction of these events in the regions within the combined systematic and statistical uncertainties on the Monte Carlo predictions discussed above. The total uncertainty on the QCD background estimation is 3.4%.

Acceptance


The theoretical uncertainty on the geometric and kinematic acceptance factor is dominated by the limited knowledge of the proton PDFs and the modeling of boson production at the LHC.

The uncertainty resulting from the choice of the PDF set is evaluated by comparing the acceptance obtained with different PDF sets (the default MRST LO*, CTEQ6.6 and HERAPDF 1.0 ref:pdf ()) and within one PDF set by re-weighting the default sample to the different error eigenvectors available for the CTEQ6.6 NLO PDF CTEQ66 (). The uncertainty is 1.6 % and 1.0%, respectively, which combines to 1.9%.

The uncertainty on the modeling of production was evaluated by comparing the default sample acceptance to that obtained from an MC@NLO sample where the parton shower is modeled by HERWIG. The difference in acceptance is found to be smaller than 0.5%.

9 Results

The results of the analysis relevant to the cross section measurement are summarized in Table 4.

2335
127  9
284  43
0.0975  0.0019
0.0799  0.0107
Table 4: Resulting numbers for the cross section calculation. The errors include statistical and systematic uncertainties here.

Within the acceptance region defined in Table 2 they translate into a fiducial cross section of:

and a total cross section of:

After correcting the cross section for the hadronic decay branching ratio = 0.6479 0.0007 ref:xsec:PDG () this yields the following inclusive cross section :

The measured cross section is in good agreement with the theoretical NNLO cross section  nb ref:FEZW (); ref:MSTW2008 (); Aad:2010yt () and the ATLAS measurements of the and cross sections ATLAS-CONF-2011-041 (); WZnewpaper (). The comparison of the cross section measurements for the different lepton final states and the theoretical expectation is shown in Figure 4. This is the first cross section measurement performed at the LHC.

Figure 4: Cross sections for the different channels measured in ATLAS with 2010 data (points). Systematic, luminosity and statistical uncertainties are added in quadrature. The theoretical NNLO expectation is also shown (dashed line), together with its uncertainty (filled area).

10 Acknowledgements

We thank CERN for the very successful operation of the LHC, as well as the support staff from our institutions without whom ATLAS could not be operated efficiently.

We acknowledge the support of ANPCyT, Argentina; YerPhI, Armenia; ARC, Australia; BMWF, Austria;
ANAS, Azerbaijan; SSTC, Belarus; CNPq and FAPESP, Brazil; NSERC, NRC and CFI, Canada; CERN; CONICYT, Chile; CAS, MOST and NSFC, China; COLCIENCIAS, Colombia; MSMT CR, MPO CR and VSC CR, Czech Republic; DNRF, DNSRC and Lundbeck Foundation, Denmark; ARTEMIS, European Union; IN2P3-CNRS, CEA-DSM/IRFU, France; GNAS, Georgia; BMBF, DFG, HGF, MPG and AvH Foundation, Germany; GSRT, Greece; ISF, MINERVA, GIF, DIP and Benoziyo Center, Israel; INFN, Italy; MEXT and JSPS, Japan; CNRST, Morocco; FOM and NWO, Netherlands; RCN, Norway; MNiSW, Poland; GRICES and FCT, Portugal; MERYS (MECTS), Romania; MES of Russia and ROSATOM, Russian Federation; JINR; MSTD, Serbia; MSSR, Slovakia; ARRS and MVZT, Slovenia; DST/NRF, South Africa; MICINN, Spain; SRC and Wallenberg Foundation, Sweden; SER, SNSF and Cantons of Bern and Geneva, Switzerland; NSC, Taiwan; TAEK, Turkey; STFC, the Royal Society and Leverhulme Trust, United Kingdom; DOE and NSF, United States of America.

The crucial computing support from all WLCG partners is acknowledged gratefully, in particular from CERN and the ATLAS Tier-1 facilities at TRIUMF (Canada), NDGF (Denmark, Norway, Sweden), CC-IN2P3 (France), KIT/GridKA (Germany), INFN-CNAF (Italy),
NL-T1 (Netherlands), PIC (Spain), ASGC (Taiwan), RAL (UK) and BNL (USA) and in the Tier-2 facilities worldwide.

References

  • (1) The ATLAS Collaboration, CERN-OPEN-2008-020, 2008.
  • (2) The ATLAS Collaboration, arXiv:1107.5003 [hep-ex].
  • (3) The ATLAS Collaboration, ATLAS Note ATL-PHYS-PUB-2010-006 (2010) .
  • (4) The ATLAS Collaboration, ATLAS Conference Note ATLAS-CONF-2011-041 (2011) .
  • (5) The ATLAS Collaboration, arXiv:1109.5141 [hep-ex].
  • (6) C. Anastasiou and L. Dixon and K. Melnikov and F. Petriello, Phys. Rev. D69 (2004) 094008.
  • (7) A. D. Martin, W. J. Stirling, R. S. Thorne and G. Watt, Eur. Phys. J. C63 (2009) 189–285.
  • (8) The ATLAS Collaboration, JHEP 12 (2010) 060.
  • (9) The UA1 Collaboration, C. Albajar et al., Phys. Lett. B 185 (1987) no. 1-2, 233 – 240.
  • (10) The CDF Collaboration, F. Abe et al., Phys. Rev. Lett 68 (1992) no. 23, 3398.
  • (11) The D0 Collaboration, B. Abbott et al., Phys. Rev. Lett. 84 (2000) 5710–5715.
  • (12) The ATLAS Collaboration, JINST 3 (2008) S08003.
  • (13) The ATLAS Collaboration, The European Physical Journal C - Particles and Fields 71 (2011) 1–37.
  • (14) The ATLAS Collaboration, ATLAS Conference Note ATLAS-CONF-2011-011 (2011) .
  • (15) T. Sjöstrand, S. Mrenna and P. Skands, JHEP 05 (2006) 026.
  • (16) A. Sherstnev and R.S. Thorne, The European Physical Journal C - Particles and Fields 55 (2008) 553–575.
  • (17) S. Frixione and B.R. Webber, JHEP 06 (2002) 029.
  • (18) G. Corcella et al., JHEP 01 (2001) 010.
  • (19) J. M. Butterworth, J. R. Forshaw and M. H. Seymour, Z. Phys. C72 (1996) 637.
  • (20) S. Jadach, J. H. Kuhn and Z. Was, Comput. Phys. Commun. 64 (1990) 275.
  • (21) E. Barberio, B. v. Eijk and Z. Was, Comput. Phys. Commun. 66 (1991) 115.
  • (22) The ATLAS Collaboration, ATLAS Note ATL-PHYS-PUB-2010-014 (2010) .
  • (23) The GEANT4 Collaboration, S. Agostinelli et al., Nucl. Instrum. Meth. A506 (2003) 250.
  • (24) The ATLAS Collaboration, Eur. Phys. J. C70 (2010) 823.
  • (25) The ATLAS Collaboration, arXiv:1110.3174 [hep-ex].
  • (26) The ATLAS Collaboration, ATLAS Conference Note ATLAS-CONF-2011-063 (2011) .
  • (27) M. Cacciari, G. P. Salam and G. Soyez, JHEP 04 (2008) 063.
  • (28) The ATLAS Collaboration, Eur. Phys. J. C71 (2011) 1512.
  • (29) T. Barillari et al., ATLAS note ATL-LARG-PUB-2009-001 (2009) .
  • (30) The ATLAS Collaboration, ATLAS Conference Note ATLAS-CONF-2011-077 (2011) .
  • (31) Y. Freund and R. Shapire in Proceedings 13th International Conference on Machine Learning. 1996.
  • (32) The ATLAS Collaboration, JHEP 1009 (2010) 056.
  • (33) The ATLAS Collaboration, ATLAS Conference Note ATLAS-CONF-2010-057 (2010) .
  • (34) The ATLAS Collaboration, ATLAS Conference Note ATLAS-CONF-2011-080 (2011) .
  • (35) The ATLAS Collaboration, ATLAS Conference Note ATLAS-CONF-2011-040 (2011) .
  • (36) The H1 and ZEUS collaborations, Journal of High Energy Physics 2010 (2010) 1–63.
  • (37) P. M. Nadolsky et al., Phys. Rev. D 78 (2008) 013004.
  • (38) Particle Data Group Collaboration, K. Nakamura et al., J. Phys. G 37 (2010) .

The ATLAS Collaboration

G. Aad, B. Abbott, J. Abdallah, A.A. Abdelalim, A. Abdesselam, O. Abdinov, B. Abi, M. Abolins, H. Abramowicz, H. Abreu, E. Acerbi, B.S. Acharya, D.L. Adams, T.N. Addy, J. Adelman, M. Aderholz, S. Adomeit, P. Adragna, T. Adye, S. Aefsky, J.A. Aguilar-Saavedra, M. Aharrouche, S.P. Ahlen, F. Ahles, A. Ahmad, M. Ahsan, G. Aielli, T. Akdogan, T.P.A. Åkesson, G. Akimoto, A.V. Akimov , A. Akiyama, M.S. Alam, M.A. Alam, J. Albert, S. Albrand, M. Aleksa, I.N. Aleksandrov, F. Alessandria, C. Alexa, G. Alexander, G. Alexandre, T. Alexopoulos, M. Alhroob, M. Aliev, G. Alimonti, J. Alison, M. Aliyev, P.P. Allport, S.E. Allwood-Spiers, J. Almond, A. Aloisio, R. Alon, A. Alonso, M.G. Alviggi, K. Amako, P. Amaral, C. Amelung, V.V. Ammosov, A. Amorim, G. Amorós, N. Amram, C. Anastopoulos, L.S. Ancu, N. Andari, T. Andeen, C.F. Anders, G. Anders, K.J. Anderson, A. Andreazza, V. Andrei, M-L. Andrieux, X.S. Anduaga, A. Angerami, F. Anghinolfi, N. Anjos, A. Annovi, A. Antonaki, M. Antonelli, A. Antonov, J. Antos, F. Anulli, S. Aoun, L. Aperio Bella, R. Apolle, G. Arabidze, I. Aracena, Y. Arai, A.T.H. Arce, J.P. Archambault, S. Arfaoui, J-F. Arguin, E. Arik, M. Arik, A.J. Armbruster, O. Arnaez, C. Arnault, A. Artamonov, G. Artoni, D. Arutinov, S. Asai, R. Asfandiyarov, S. Ask, B. Åsman, L. Asquith, K. Assamagan, A. Astbury, A. Astvatsatourov, G. Atoian, B. Aubert, E. Auge, K. Augsten, M. Aurousseau, N. Austin, G. Avolio, R. Avramidou, D. Axen, C. Ay, G. Azuelos, Y. Azuma, M.A. Baak, G. Baccaglioni, C. Bacci, A.M. Bach, H. Bachacou, K. Bachas, G. Bachy, M. Backes, M. Backhaus, E. Badescu, P. Bagnaia, S. Bahinipati, Y. Bai, D.C. Bailey, T. Bain, J.T. Baines, O.K. Baker, M.D. Baker, S. Baker, E. Banas, P. Banerjee, Sw. Banerjee, D. Banfi, A. Bangert, V. Bansal, H.S. Bansil, L. Barak, S.P. Baranov, A. Barashkou, A. Barbaro Galtieri, T. Barber, E.L. Barberio, D. Barberis, M. Barbero, D.Y. Bardin, T. Barillari, M. Barisonzi, T. Barklow, N. Barlow, B.M. Barnett, R.M. Barnett, A. Baroncelli, G. Barone, A.J. Barr, F. Barreiro, J. Barreiro Guimarães da Costa, P. Barrillon, R. Bartoldus, A.E. Barton, D. Bartsch, V. Bartsch, R.L. Bates, L. Batkova, J.R. Batley, A. Battaglia, M. Battistin, G. Battistoni, F. Bauer, H.S. Bawa, B. Beare, T. Beau, P.H. Beauchemin, R. Beccherle, P. Bechtle, H.P. Beck, M. Beckingham, K.H. Becks, A.J. Beddall, A. Beddall, S. Bedikian, V.A. Bednyakov, C.P. Bee, M. Begel, S. Behar Harpaz, P.K. Behera, M. Beimforde, C. Belanger-Champagne, P.J. Bell, W.H. Bell, G. Bella, L. Bellagamba, F. Bellina, M. Bellomo, A. Belloni, O. Beloborodova, K. Belotskiy, O. Beltramello, S. Ben Ami, O. Benary, D. Benchekroun, C. Benchouk, M. Bendel, N. Benekos, Y. Benhammou, D.P. Benjamin, M. Benoit, J.R. Bensinger, K. Benslama, S. Bentvelsen, D. Berge, E. Bergeaas Kuutmann, N. Berger, F. Berghaus, E. Berglund, J. Beringer, K. Bernardet, P. Bernat, R. Bernhard, C. Bernius, T. Berry, A. Bertin, F. Bertinelli, F. Bertolucci, M.I. Besana, N. Besson, S. Bethke, W. Bhimji, R.M. Bianchi, M. Bianco, O. Biebel, S.P. Bieniek, K. Bierwagen, J. Biesiada, M. Biglietti, H. Bilokon, M. Bindi, S. Binet, A. Bingul, C. Bini, C. Biscarat, U. Bitenc, K.M. Black, R.E. Blair, J.-B. Blanchard, G. Blanchot, T. Blazek, C. Blocker, J. Blocki, A. Blondel, W. Blum, U. Blumenschein, G.J. Bobbink, V.B. Bobrovnikov, S.S. Bocchetta, A. Bocci, C.R. Boddy, M. Boehler, J. Boek, N. Boelaert, S. Böser, J.A. Bogaerts, A. Bogdanchikov, A. Bogouch, C. Bohm, V. Boisvert, T. Bold, V. Boldea, N.M. Bolnet, M. Bona, V.G. Bondarenko, M. Bondioli, M. Boonekamp, G. Boorman, C.N. Booth, S. Bordoni, C. Borer, A. Borisov, G. Borissov, I. Borjanovic, S. Borroni, K. Bos, D. Boscherini, M. Bosman, H. Boterenbrood, D. Botterill, J. Bouchami, J. Boudreau, E.V. Bouhova-Thacker, C. Bourdarios, N. Bousson, A. Boveia, J. Boyd, I.R. Boyko, N.I. Bozhko, I. Bozovic-Jelisavcic, J. Bracinik, A. Braem, P. Branchini, G.W. Brandenburg, A. Brandt, G. Brandt, O. Brandt, U. Bratzler, B. Brau, J.E. Brau, H.M. Braun, B. Brelier, J. Bremer, R. Brenner, S. Bressler, D. Breton, D. Britton, F.M. Brochu, I. Brock, R. Brock, T.J. Brodbeck, E. Brodet, F. Broggi, C. Bromberg, G. Brooijmans, W.K. Brooks, G. Brown, H. Brown, P.A. Bruckman de Renstrom, D. Bruncko, R. Bruneliere, S. Brunet, A. Bruni, G. Bruni, M. Bruschi, T. Buanes, F. Bucci, J. Buchanan, N.J. Buchanan, P. Buchholz, R.M. Buckingham, A.G. Buckley, S.I. Buda, I.A. Budagov, B. Budick, V. Büscher, L. Bugge, D. Buira-Clark, O. Bulekov, M. Bunse, T. Buran, H. Burckhart, S. Burdin, T. Burgess, S. Burke, E. Busato, P. Bussey, C.P. Buszello, F. Butin, B. Butler, J.M. Butler, C.M. Buttar, J.M. Butterworth, W. Buttinger, T. Byatt, S. Cabrera Urbán, D. Caforio, O. Cakir, P. Calafiura, G. Calderini, P. Calfayan, R. Calkins, L.P. Caloba, R. Caloi, D. Calvet, S. Calvet, R. Camacho Toro, P. Camarri, M. Cambiaghi, D. Cameron, S. Campana, M. Campanelli, V. Canale, F. Canelli, A. Canepa, J. Cantero, L. Capasso, M.D.M. Capeans Garrido, I. Caprini, M. Caprini, D. Capriotti, M. Capua, R. Caputo, C. Caramarcu, R. Cardarelli, T. Carli, G. Carlino, L. Carminati, B. Caron, S. Caron, G.D. Carrillo Montoya, A.A. Carter, J.R. Carter, J. Carvalho, D. Casadei, M.P. Casado, M. Cascella, C. Caso, A.M. Castaneda Hernandez, E. Castaneda-Miranda, V. Castillo Gimenez, N.F. Castro, G. Cataldi, F. Cataneo, A. Catinaccio, J.R. Catmore, A. Cattai, G. Cattani, S. Caughron, D. Cauz, P. Cavalleri, D. Cavalli, M. Cavalli-Sforza, V. Cavasinni, F. Ceradini, A.S. Cerqueira, A. Cerri, L. Cerrito, F. Cerutti, S.A. Cetin, F. Cevenini, A. Chafaq, D. Chakraborty, K. Chan, B. Chapleau, J.D. Chapman, J.W. Chapman, E. Chareyre, D.G. Charlton, V. Chavda, C.A. Chavez Barajas, S. Cheatham, S. Chekanov, S.V. Chekulaev, G.A. Chelkov, M.A. Chelstowska, C. Chen, H. Chen, S. Chen, T. Chen, X. Chen, S. Cheng, A. Cheplakov, V.F. Chepurnov, R. Cherkaoui El Moursli, V. Chernyatin, E. Cheu, S.L. Cheung, L. Chevalier, G. Chiefari, L. Chikovani, J.T. Childers, A. Chilingarov, G. Chiodini, M.V. Chizhov, G. Choudalakis, S. Chouridou, I.A. Christidi, A. Christov, D. Chromek-Burckhart, M.L. Chu, J. Chudoba, G. Ciapetti, K. Ciba, A.K. Ciftci, R. Ciftci, D. Cinca, V. Cindro, M.D. Ciobotaru, C. Ciocca, A. Ciocio, M. Cirilli, M. Ciubancan, A. Clark, P.J. Clark, W. Cleland, J.C. Clemens, B. Clement, C. Clement, R.W. Clifft, Y. Coadou, M. Cobal, A. Coccaro, J. Cochran, P. Coe, J.G. Cogan, J. Coggeshall, E. Cogneras, C.D. Cojocaru, J. Colas, A.P. Colijn, C. Collard, N.J. Collins, C. Collins-Tooth, J. Collot, G. Colon, P. Conde Muiño, E. Coniavitis, M.C. Conidi, M. Consonni, V. Consorti, S. Constantinescu, C. Conta, F. Conventi, J. Cook, M. Cooke, B.D. Cooper, A.M. Cooper-Sarkar, N.J. Cooper-Smith, K. Copic, T. Cornelissen, M. Corradi, F. Corriveau, A. Cortes-Gonzalez, G. Cortiana, G. Costa, M.J. Costa, D. Costanzo, T. Costin, D. Côté, L. Courneyea, G. Cowan, C. Cowden, B.E. Cox, K. Cranmer, F. Crescioli, M. Cristinziani, G. Crosetti, R. Crupi, S. Crépé-Renaudin, C.-M. Cuciuc, C. Cuenca Almenar, T. Cuhadar Donszelmann, M. Curatolo, C.J. Curtis, P. Cwetanski, H. Czirr, Z. Czyczula, S. D’Auria, M. D’Onofrio, A. D’Orazio, P.V.M. Da Silva, C. Da Via, W. Dabrowski, T. Dai, C. Dallapiccola, M. Dam, M. Dameri, D.S. Damiani, H.O. Danielsson, D. Dannheim, V. Dao, G. Darbo, G.L. Darlea, C. Daum, J.P. Dauvergne , W. Davey, T. Davidek, N. Davidson, R. Davidson, E. Davies, M. Davies, A.R. Davison, Y. Davygora, E. Dawe, I. Dawson, J.W. Dawson, R.K. Daya, K. De