Measurement of the ttbar production cross section in pp collisions at sqrt(s) = 8 TeV in dilepton final states containing one tau lepton

The top-quark pair production cross section is measured in final states with one electron or muon and one hadronically decaying tau lepton from the process ttbar to (l nu[l]) (tau nu[tau]) bbbar, where l = e, mu. The data sample corresponds to an integrated luminosity of 19.6 inverse femtobarns collected with the CMS detector in proton-proton collisions at sqrt(s) = 8 TeV. The measured cross section sigma[ttbar] = 257 +/- 3 (stat) +/- 24 (syst) +/- 7 (lumi) pb, assuming a top-quark mass of 172.5 GeV, is consistent with the standard model prediction.


Introduction
Top quarks at the CERN LHC are mostly produced in pairs with the subsequent decays tt → W + bW − b.The decay modes of the two W bosons determine the event signature.The dilepton decay channel corresponds to the case in which both W bosons decay into leptons, where the term lepton usually refers to electrons or muons, as studied in Refs.[1,2].In this letter we measure the production cross section of top-quark pairs by considering dilepton decays where one W boson promptly decays into ν , with = e or µ, and the other decays into τν τ , tt → ( ν )(τν τ )bb.The expected fraction of these events is 4/81 of all tt decays.The τ lepton is identified by means of its hadronic decay products, with a branching fraction B(τ → hadrons + ν τ ) 65%, to produce a narrow jet with a small number of charged hadrons, denoted as τ h .The cross section is measured by counting the number of τ h + X events consistent with originating from tt production, after subtracting the contributions from other processes, and correcting for the efficiency of the event selection.A similar method was used in pp collisions at a centre-ofmass energy of √ s = 7 TeV [3].This "τ dilepton" channel is of particular interest because it is a natural background process to the search for a charged Higgs boson [4,5] with a mass smaller than that of the top quark.In this case, the production chain tt → H + bW − b, with H + → τ + ν τ , could give rise to differences with respect to the standard model (SM) prediction of the number of tt events with a τ lepton [6].The present measurement is based on data collected by the CMS experiment in pp collisions at √ s = 8 TeV corresponding to an integrated luminosity of 19.6 fb −1 .The relative accuracy of the measurement improves over previous results [7][8][9][10][11].
The CMS detector is briefly introduced in Section 2, followed by details of the simulated samples in Section 3, and a brief description of the event reconstruction and event selection in Section 4. The descriptions of the background determination and the systematic uncertainties are given in Sections 5 and 6, respectively.The measurement of the cross section is discussed in Section 7, and the results are summarised in Section 8.

The CMS detector
The central feature of the CMS apparatus is a superconducting solenoid of 6 m internal diameter and 13 m in length, providing a magnetic field of 3.8 T. Within the superconducting solenoid volume are a silicon pixel and strip tracker, a lead tungstate crystal electromagnetic calorimeter, and a brass/scintillator hadron calorimeter, each composed of a barrel and two endcap sections.The calorimetry provides high-resolution energy and direction measurements of electrons and hadronic jets.Muons are identified using gas-ionization detectors embedded in the steel flux-return yoke outside the solenoid.Extensive forward calorimetry complements the coverage provided by the barrel and endcap detectors.The CMS experiment uses a righthanded coordinate system, with the origin at the nominal interaction point, the x axis pointing to the centre of the LHC ring, the y axis pointing up (perpendicular to the LHC plane), and the z axis along the anticlockwise-beam direction.The polar angle θ is measured from the positive z axis and the azimuthal angle ϕ is measured in the x-y plane.Charged particle trajectories are measured covering 0 < ϕ ≤ 2π in azimuth and |η| < 2.5, where the pseudorapidity η is defined as η = − ln[tan(θ/2)].The detector is nearly hermetic, allowing for energy balance measurements in the plane transverse to the beam directions.A two-level trigger system selects the most interesting proton-proton collision events for use in physics analyses.A more detailed description of the CMS detector can be found elsewhere [12].

Data and simulation samples
Events are selected online by a trigger requiring a single isolated electron (muon) with transverse momentum p T > 27 (24) GeV and |η| < 2.5 (2.1).
This measurement makes use of simulated samples of tt events as well as other processes that mimic the τ h decay signature.These samples are used to optimise the event selection, to calculate the acceptance for tt events, and to estimate some of the backgrounds in the analysis.
The signal acceptance and tt dilepton background are evaluated using a version of MADGRAPH which includes the effects of spin correlations [13,14].The number of expected tt events is estimated with the approximate next-to-next-to-leading-order (NNLO) SM cross section of 251.7 +6.3 −8.6 (scale) ± 6.5 (PDF) pb [15], for a top-quark mass of 172.5 GeV, where the first uncertainty is due to renormalisation and factorisation scales, and the second is due to the choice of parton distribution functions (PDFs).The generated events are subsequently processed with PYTHIA 6.426 [16] which performs the hadronisation of partons.Soft radiation is matched to the contributions from direct emissions accounted for in the matrix-element calculations using the k T -MLM approach [17].The τ lepton decays are simulated using TAUOLA 27.121.5 [18], which accounts for the τ-lepton polarization.
Simulated events are processed using the full CMS detector simulation based on GEANT4 [27,28], followed by a detailed trigger emulation and event reconstruction.For both signal and background events, additional pp interactions (pileup) in the same or nearby bunch crossings are simulated with PYTHIA and superimposed on the hard collision, using a pileup multiplicity distribution that reflects the luminosity profile of the analysed data.

Event selection
Events are reconstructed with the particle-flow (PF) algorithm [29,30], which combines information from all sub-detectors to identify and reconstruct individual electrons, muons, photons, charged and neutral hadrons.Electrons are identified with a multivariate discriminant combining several quantities describing the track quality, the shape of the energy deposits in the electromagnetic calorimeter, and the compatibility of the measurements from the tracker and the electromagnetic calorimeter [31].Muons are identified with additional requirements on the quality of the track reconstruction and on the number of measurements in the tracker and the muon systems [32].The PF algorithm uses the charged particles originating from the primary collision vertex, chosen as the reconstructed vertex with the largest ∑ p 2 T of the associated tracks.These charged particles and all neutral particles then provide the input to the anti-k T jet clustering algorithm with a distance parameter of 0.5 [33].The jet momentum is determined from the vector sum of particle momenta in the jet.After jet energies are corrected for additional pileup contributions and for detector effects, they are found in simulations to be within 5-10% of the actual jet momentum [34].The missing transverse energy E miss T is calculated as the magnitude of the vector sum of momenta from all reconstructed particles in the plane transverse to the beam.
In addition, higher-level observables such as b-tagging discriminators and isolation variables are determined.The lepton relative isolation is defined as I rel = (E T, ch + E T, nh + E T, ph )/p T , where E T, ch , E T, nh , and E T, ph refer to the transverse energy deposited by charged hadrons, neutral hadrons, and photons, respectively, in a cone of radius R = √ (∆ϕ) 2 + (∆η) 2 = 0.4 around the lepton track, whose transverse momentum is p T .An electron (muon) candidate is considered to be non-isolated and is rejected if I rel > 0.1 (>0.12).
The hadronic products of the τ-lepton decay are reconstructed using a jet as the initial seed, and are then classified as having one or three charged hadrons with the "hadron-plus-strips" algorithm [35,36].Calorimeter energy deposits clustered along strips in the ϕ direction are used for neutral pion identification.Then, the decay modes, four-momenta, and isolation quantities of the τ h are determined, and the following categories are considered: single hadron, hadron plus a strip, hadron plus two strips, and three hadrons.These categories together encompass approximately 95% of hadronic τ-lepton decays.The sum of the charged hadron charges provides the τ h charge.The τ h -jet momentum is required to match the direction of the original jet within a maximum distance R = 0.1.Isolation criteria require that there be no additional charged hadrons with p T > 1.0 GeV or photons with transverse energy E T > 1.5 GeV within a cone of size R = 0.5 around the direction of the τ h jet.Electrons and muons misidentified as τ h are suppressed using algorithms that combine information from the tracker, calorimeters, and muon detectors [12].The τ h identification efficiency, defined as the ratio of the number of selected τ h candidates divided by the number of hadronic τ-lepton decays in tt events, is approximately 50% for p τ h T > 20 GeV, with a probability of approximately 1% for generic jets to be misidentified as a τ h jet.
The combined secondary vertex (CSV) algorithm [37] is used to identify jets originating from the hadronisation of b quarks.The algorithm combines the information about track impact parameters and secondary vertices within jets into a likelihood discriminant to provide separation between b jets and jets originating from light quarks, gluons, or charm quarks.The output of this CSV discriminant has values between zero and one; a jet with a CSV value above a certain threshold is referred to as being "b tagged".We choose a working point where the b-tagging efficiency is approximately 60%, as measured in a data sample of events enriched with jets from semileptonic b-hadron decays.The misidentification rate of light-flavour jets is estimated from inclusive jet studies and is measured to be about 0.1% for jets with p T > 30 GeV.
Events are preselected by requiring exactly one isolated electron (muon) with transverse momentum p T > 35 (30) GeV and |η| < 2. 5 (2.1), at least two jets with p T > 30 GeV, and one additional jet with p T > 20 GeV.The selected jets must be within |η| < 2.4.The electron or muon is required to be separated from any jet in the (η, ϕ) plane by a distance R > 0.4.Events with any additional loosely isolated, I rel < 0.2, electron (muon) of p T > 15 (10) GeV are rejected.Further event selection requirements include E miss T > 40 GeV and only one τ h with p T > 20 GeV and |η| < 2.4.The τ h and the lepton are required to have electric charges of opposite sign (OS).At least one of the jets is required to be identified as originating from b-quark hadronisation (b-tagged).
Figure 1 shows, for the sum of the eτ h and µτ h final states, a comparison between data and simulation of the number of b-tagged jets in each event N b-tag after all the selection criteria have been applied.The distributions of the τ h p T and E miss T after the final event selection are shown in the leftand rightpanels of Fig. 2, respectively.The distributions show agreement between the observed numbers of events and the expected numbers of signal and background events obtained from the simulation.Following the final section, additional kinematic features of the tt events are studied to evaluate the agreement between the observed data and the predicted sum of signal and background.For each event, two combinations of the invariant mass are reconstructed by pairing the τ h with the two candidate b-tagged jets: (1) in events with two or more b-tagged jets, the two combinations are based on the two b-tagged jets with the highest value of the discriminator; (2) in events with one b-tagged jet, this is used for the first combination, while the non-b-tagged jet with the highest p T is used to form the second combination.For the two combinations, the invariant mass with the lowest value is shown in Fig. 3 (left), for the eτ h and µτ h channels combined.
For each event, the top-quark mass m top is reconstructed using the KINb algorithm [38,39].As there are multiple neutrinos in each event and the identification of which leptons and jets are associated with which W boson is unknown, the reconstruction of m top leads to an underconstrained system.The KINb algorithm applies constraints on the W boson mass, the mass difference between the top and anti-top quark, and the longitudinal momentum of the dilepton system.For each event, solutions to the kinematic equations are evaluated, varying the jet momenta and the direction of E miss T within their resolutions.For each set of variations and each lepton-jet combination, the kinematic equations allow up to four solutions; the one with the lowest tt invariant mass is accepted if the mass difference between the two top quarks is less than 3 GeV.For each event, the accepted solutions corresponding to the two possible lepton-jet combinations are counted and the combination with the largest number of solutions is chosen and m top is obtained by fitting the peak of this distribution.The events in which solutions are found are shown in Fig. 3

Background estimate
The main background (misidentified τ h ) comes from events with one lepton (electron or muon), significant E miss T , and three or more jets, where one jet is misidentified as a τ h jet [6].The dominant source is tt lepton+jet events.The misidentified τ h background accounts also for events with W bosons produced in association with jets, either genuine W+jet or single-topquark production, and for QCD multijet events.In order to estimate this background from data, the misidentification probability w(jet → τ h ) is parameterised as a function of the jet p T , η, and width (R jet ).The quantity R jet is defined as , where σ η (σ ϕ ) expresses the extent in η (ϕ) of the jet cluster [34].
The probability w(jet → τ h ) is evaluated from two control samples: • w W+jets : from a W+jet event sample, selected by requiring one isolated muon with p T > 20 GeV and |η| < 2.1, and at least one jet with p T > 20 GeV and |η| < 2.4; • w QCD : from a QCD multijet sample, triggered by one jet with p T > 40 GeV, selected by requiring events to have at least two jets with p T > 20 GeV and |η| < 2.4, where the triggering jet is removed from the misidentification rate calculation to avoid a trigger bias.
Both probabilities are evaluated in simulated events as well as in data, with good agreement found between the results from simulation and data [35].
The number of events containing misidentified τ h candidates is then determined as where j is the jet index of event i, and m is the number of jets in each event and M is the total number of events.The quantity N other is the expected 20% contamination from signal and other processes to the misidentified background as estimated from simulated samples.The value of N other is evaluated by applying the procedure described above to simulated events of Z/γ * → ττ, single-top-quark production, diboson production, and the tt processes included in the misidentified τ h background estimation.
Jets in QCD multijet events originate mainly from gluons, while in W+jet events they are predominantly from quarks.The quark and gluon composition in the misidentified τ h events lies between these two control samples.As w QCD < w W+jets , the actual N misid value is under-(over-) estimated by applying the w QCD (w W+jets ) probability.We determine from simulation the rate for the misidentification of a quark or gluon jet to be identified as a τ h and the quark/gluon composition in the W+jet and multijet samples.From these quantities we derive the following combination: where the misidentification rates extracted from the data control samples discussed above, are combined with the scale factors SF determined from simulation: SF QCD = 0.83 and SF W+jet = 0.17.The corresponding systematic uncertainty is obtained from Eq. ( 2) by weighting the relative deviations of N misid W+jet and N misid QCD from N misid with the related scale factors.This results in an uncertainty of 7% for both eτ h and µτ h channels.
The efficiency of the OS requirement ε OS is determined from simulated lepton+jet tt events and is applied in order to obtain the misidentified τ h background after the final event selection N misid OS , where N misid OS = ε OS • N misid .We find values of ε OS = 0.729 ± 0.002 (stat) ± 0.004 (syst) for the eτ h selection and ε OS = 0.731 ± 0.002 (stat) ± 0.003 (syst) for the µτ h selection, where all sources of systematic uncertainty are accounted for in the modelling of the simulated tt lepton+jet events.

Systematic uncertainties
The following systematic uncertainties in the signal reconstruction efficiency, background determination, and theoretical assumptions on the tt production are considered, where relative values refer to the cross section uncertainty unless explicitly stated otherwise: τ h identification and energy scale -the uncertainty in the τ h reconstruction accounts for the identification efficiency (6%) and the τ h jet energy scale that results in 2.4% (2.5%) for the eτ h (µτ h ) channel is estimated by varying the p T of the τ h jet by 3% [35,36]; misidentified τ h -the uncertainty related to the misidentified τ h background process, discussed in Section 5, is obtained by propagating the 7% uncertainty on N misid to the cross section determination and results in 4.3% for both channels.It also includes the uncertainty in the OS efficiency determination; b-jet tagging, light-flavour jet misidentification -the reconstruction of a light flavour jet as a b quark is defined as mistagging.The uncertainty due to b (mis)tagging is estimated to reflect the data-to-simulation scale factors and corresponding uncertainties for b-tagging and mistagging efficiencies [37].When propagated to the cross section measurement, they amount to 1.6% for both eτ h and µτ h channels; jet energy scale, jet energy resolution, E miss T -the jet energy scale (JES) uncertainty is estimated [34] by varying the jet energy within the p T -and η-dependent JES uncertainties per jet, and taking into account the uncertainty due to pileup and parton flavour.The jet energy resolution (JER) is estimated by smearing the jet energy in simulation within the η-dependent JER uncertainties per jet.The JES and JER uncertainties are propagated in order to estimate the uncertainty of the E miss T scale.In addition, modelling of the E miss T component, which is not clustered in jets, is also considered.The resulting uncertainty from propagating these effects to the cross section measurement is 1.9% for both the eτ h and µτ h channels; lepton reconstruction -the uncertainties due to trigger, lepton identification, isolation, and lepton energy scale are calculated from independent samples [31,32], and yield 0.8% (0.6%) for the eτ h (µτ h ) channel; other backgrounds -an overall 0.6% (0.7%) uncertainty for the eτ h (µτ h ) channel is due to other minor backgrounds, accounting for the uncertainties related to the theoretical cross sections, JES, and b-tagging in these simulated samples, and the → τ h ( = e, µ) misidentification in the Z/γ * → + − and tt dilepton processes; luminosity -the uncertainty of the integrated luminosity is 2.6% [40]; matrix element matching -the theoretical uncertainty due to the matrix element (ME) and parton shower (PS) matching is estimated by varying up and down by a factor of two the threshold between jet production at the ME level and via PS, and it results in 1.7% (1.3%) for the eτ h (µτ h ) channel; factorisation/renormalisation scale -the modelling uncertainty in the signal acceptance due to the factorisation and renormalisation scale choices is estimated by varying them simultaneously up and down by a factor of two from the nominal value equal to the Q 2 in the event, with an uncertainty of 2.9% found for both channels; generator -the uncertainty due to the choice of the generator is estimated as the relative difference between the acceptances evaluated with MADGRAPH and POWHEG [20][21][22]41] after the full event selection and results in 1.5%; hadronisation -the uncertainty in the hadronisation scheme is evaluated from the relative differences between the acceptances from POWHEG+PYTHIA and POWHEG+HERWIG samples, estimated prior to the b-tagging or τ h jet requirement, resulting in a 1.7% uncertainty; other sources from theory -we consider the uncertainty related to the top-quark p T scale modelling by varying the top-quark p T spectrum and evaluating the change in the signal acceptance, resulting in 0.6%, and the uncertainty related to the PDF variations following the PDF4LHC prescriptions [42], resulting in 0.7%.
All systematic uncertainties for signal and background processes are summarised in Table 1.

Cross section measurement
The number of expected signal and background events as well as the number of observed events after all selections are summarised in Table 2.The statistical and systematic uncertainties are also shown.The tt production cross section measured from τ dilepton events is σ tt = where N is the number of observed candidate events, B is the estimate of the background and L is the integrated luminosity.The total acceptance A tot is the product of the branching fractions, geometrical and kinematic acceptance, trigger, lepton identification, and the overall reconstruction efficiency.It is evaluated with respect to the inclusive tt sample.
The statistical uncertainties are due to the limited number of simulated events and the systematic uncertainties are estimated by accounting for all sources listed in Table 1.The statistical and systematic uncertainties listed in Table 2 are propagated to the final cross section measurements: σ tt (eτ h ) = 255 ± 4 (stat) ± 24 (syst) ± 7 (lumi) pb; σ tt (µτ h ) = 258 ± 4 (stat) ± 24 (syst) ± 7 (lumi) pb.
The BLUE method [43] is used to combine the cross section measurements in the eτ h and µτ h channels, with the corresponding weights of 0.47 and 0.53, respectively.Lepton reconstruction uncertainties are uncorrelated, while all other uncertainties are assumed 100% correlated.
With this method we obtain a combined result of σ tt = 257 ± 3 (stat) ± 24 (syst) ± 7 (lumi) pb, in agreement with the NNLO expectation of 251.7 +6.3 −8.6 (scales) ± 6.5 (PDF) pb [15].The dependence on the top-quark mass has been studied for the range 160-185 GeV and is well described by a linear variation.If we adjust our result to the current world average value of 173.3 GeV [44], we obtain a cross section that is lower by 3.1 pb.

Summary
A measurement of the tt production cross section in the channel tt → ( ν )(τν τ )bb is presented, where is an electron or a muon, and the τ lepton is reconstructed through its hadronic decays.The data sample corresponds to an integrated luminosity of 19.6 fb −1 collected in proton-proton collisions at √ s = 8 TeV.Events are selected by requiring the presence of one isolated electron or muon, two or more jets (at least one of which is b-tagged), significant missing transverse energy, and one τ.The largest background contribution is estimated from data and consists of tt events with one W boson decaying into jets, where one jet is misidentified as a τ.The measured cross section is σ tt = 257 ± 3 (stat) ± 24 (syst) ± 7 (lumi) pb for a top-quark mass of 172.5 GeV.This measurement improves over previous results in this decay channel, and it is in good agreement with the standard model expectation and other measurements of the tt cross section at same centre-of-mass energy.

References
[1] ATLAS Collaboration, "Measurement of the t t production cross-section using eµ events with b-tagged jets in pp collisions at √ s = 7 and 8 TeV with the ATLAS detector", (2014).arXiv:1406.5375.

Figure 1 :
Figure 1: The b-tagged jet multiplicity after the full event selection.The simulated contributions are normalised to the SM predicted values.The hatched area shows the total uncertainty.

Figure 2 :
Figure 2: Distribution of the τ h p T (left) and E miss T (right) after the full event selection, for the eτ h and µτ h channels combined.The simulated contributions are normalised to the SM predicted values.The hatched area shows the total uncertainty.The last bins include the overflow events.

Figure 3 :
Figure 3: (left) Minimum invariant mass reconstructed by pairing the τ h with either a b-jet candidate or with the highest p T non b-tagged jet, as described in the text.(right) Distribution of the reconstructed top-quark mass m top for the τ h candidate events after the full event selection.Data (points) are compared with the sum of signal and background yields, for the eτ h and µτ h channels combined.The simulated contributions are normalised to the SM predicted values.The hatched area shows the total uncertainty.The last bins include the overflow events.
(right).Data are in agreement with the expected sum of signal and background events.

Table 1 :
List of systematic uncertainties in the cross section measurement, and their combination.Lepton reconstruction uncertainties are uncorrelated, while all other uncertainties are assumed 100% correlated.

Table 2 :
Number of expected events for signal (assuming m top = 172.5 GeV) and backgrounds.The background from misidentified τ h is estimated from data, while the other backgrounds are estimated from simulation.Statistical and systematic uncertainties are shown.