Measurement of the ttbar production cross section and top quark mass extraction using dilepton events in ppbar collisions

We present a measurement of the top quark pair production cross section in ppbar collisions at \sqrt{s}=1.96 TeV using approximately 1 fb^{-1} of data collected with the D0 detector. We consider decay channels containing two high pt charged leptons where one lepton is identified as an electron or a muon while the other lepton can be an electron, a muon or a hadronically decaying tau lepton. For a mass of the top quark of 170 GeV, the measured cross section is 7.5 +1.0-1.0 (stat) +0.7-0.6 (syst) +0.6-0.5 (lumi) pb. Using lepton+tau events only, we measure: \sigma_ttbar \times B(ttbar to ltau bbbar) = 0.13 +0.09-0.08 (stat) +0.06-0.06 (syst) +0.02-0.02 (lumi) pb. Comparing the measured cross section as a function of the mass of the top quark with a partial next-to-next-to leading order Quantum Chromodynamics theoretical prediction, we extract a mass of the top quark of 171.5 +9.9-8.8 GeV, in agreement with direct measurements.

(Dated: July 14, 2009) We present a measurement of the top quark pair production cross section in pp collisions at √ s = 1.96TeV using approximately 1 fb −1 collected with the D0 detector.We consider decay channels containing two high pT charged leptons where one lepton is identified as an electron or a muon while the other lepton can be an electron, a muon or a hadronically decaying τ lepton.For a mass of the top quark of 170 GeV, the measured cross section is 7.5 +1.0 −1.0 (stat) +0.7 −0.6 (syst) +0.6 −0.5 (lumi) pb.Using ℓτ events only, we measure: σ t t × B(t t → ℓτ b b) = 0.13 +0.09  −0.08 (stat) +0.06 −0.06 (syst) +0.02 −0.02 (lumi) pb.Comparing the measured cross section as a function of the mass of the top quark with a partial next-to-next-to leading order Quantum Chromodynamics theoretical prediction, we extract a mass of the top quark of 171.5 +9.9 −8. 8 GeV, in agreement with direct measurements.
PACS numbers: 14.65.Ha, 13.85.Lg, 13.85.Qk, 14.60Fg, 12.15.Ff The top quark, first observed at Fermilab in 1995 [1,2], is the heaviest known elementary particle.In many extensions of the standard model (SM) new physics is predicted in connection with top quarks.In the SM, top quarks are predicted to decay into a W boson and a b quark with a branching fraction of nearly 100% [3].For approximately 10% of all top-antitop quark (t t) events, both W bosons decay leptonically and generate final states containing two leptons [3].In addition, these final states are characterized by the presence of two high energy jets resulting from hadronization of the two b quarks and large imbalance in transverse momentum ( E T ) due to several undetected neutrinos from the W boson decays.
New physics in the production or decay of the top quark may lead to significant deviations in the measured t t cross section (σ t t) from the SM prediction.Since new physics could have a different impact on different final states, it is important to measure σ t t precisely in all possible decay channels.Channels including a τ lepton in the final state are of particular interest, since the decay chain involves only third generation fermions.Owing to the significant dependence of the t t cross section on the mass of the top quark (m t ), a precise cross section measurement allows the extraction of the mass of the top quark in a way complementary to direct reconstruction methods and hence provides a valuable consistency check.A measurement of the mass of the top quark is important since together with that of the W boson, it allows one to place indirect constraints on the mass of the SM Higgs boson.
In this Letter, we present a measurement of σ t t using approximatively 1 fb −1 from Run II of the Fermilab Tevatron pp collider operated at √ s = 1.96TeV, and collected with the D0 detector.We consider dilepton final states with two identified electrons or muons from the W boson leptonic decays, i.e., ee, eµ and µµ, and final states with a τ lepton that decays into hadrons+ν τ from the decay of one W boson and an accompanying electron or muon from the other W boson, i.e., eτ and µτ .Throughout the text, these final states will be referred to as ℓℓ and ℓτ channels, respectively.Dilepton channels also have contributions from events where both τ leptons decay into electrons or muons.Previous measurements of σ t t in the dilepton channel were reported in [4,5].We update the D0 measurement [4] using more integrated luminosity and include the ℓτ final states in the result.We also present a measurement of σ t t × B(t t → ℓτ b b).In addition, we explore the dependence of σ t t on the mass of the top quark, and through a comparison with higher order Quantum Chromodynamics (QCD) calculations computed in the pole mass scheme, we extract a value for the mass of the top quark.
The D0 detector has a central tracking system, consisting of a silicon microstrip tracker (SMT) and a central fiber tracker, both located within a 2 T superconducting solenoidal magnet [6].A liquid argon and uranium calorimeter has a central section covering pseudorapidities |η| up to ≈ 1.1 [7], and two end calorimeters (EC) that extend coverage to |η| ≈ 4.2, with all three calorimeters housed in separate cryostats [8].An outer muon system, covering |η| < 2, consists of a layer of tracking detectors and scintillation trigger counters in front of 1.8 T iron toroids, followed by two similar layers after the toroids [9].The luminosity is measured using plastic scintillator arrays placed in front of the EC cryostats.The trigger and data acquisition systems are designed to accommodate the high luminosities of Run II.The dilepton triggers used in the ℓℓ channels are described in Ref. [4].The ℓτ channel uses triggers requiring one lepton and one jet.The trigger efficiency for signal events passing the selection and acceptance cuts varies from 78% to 98% depending on the channel.
Electrons are identified as clusters of energy deposits in calorimeter cells satisfying the following requirements: (i) the fraction of energy deposited in the electromagnetic section of the calorimeter is at least 90% of the total cluster energy, (ii) the energy is concentrated in a narrow cone, and isolated from other energy deposits, (iii) the shape of the shower is compatible with that of an electron, and (iv) a track extrapolated from the tracking system points to the cluster.To further reduce backgrounds (see below for background description) we use a likelihood discriminant that selects prompt isolated electrons, based on tracking and calorimetric information.Both central (|η| < 1.1) and forward (1.5 < |η| < 2.5) electron candidates are accepted.
Muon trajectories are reconstructed using hits in three layers of the outer muon system along with matching tracks in the inner tracker.The energy deposited within an annulus 0.1 < (∆η) 2 + (∆φ) 2 < 0.4 around the muon direction (where φ is the azimuthal angle) must be less than 15% of the muon p T , for all channels except µµ, while for the µµ final state, the selected muons must not lie within the cone of any reconstructed jet.To reduce background further, the sum of the track momenta in a cone around the muon track has to be smaller than 15% of the muon p T .Moreover, the fraction of prompt muons is increased by requiring that the distance of closest approach of the muon track to the primary vertex is small.
A hadronically decaying τ lepton is characterized by a narrow jet of low track multiplicity.The τ lepton reconstruction is seeded either by a calorimeter energy cluster using the D0 Run II cone algorithm [12] with radius R = 0.3 or by a track.Three types of τ decays are defined as (i) τ -type 1 (π ± -like), consisting of a single track, with energy deposition in the hadronic calorimeter, (ii) τtype 2 (ρ ± -like), a single track, with an energy deposit in both the hadronic and the electromagnetic calorimeters and (iii) τ -type 3, having two or three tracks, forming an invariant mass < 1.1 or < 1.7 GeV, respectively.The total sum of the particle charges for τ -type 3 is required to be ±1 or ±2.A set of neural networks (NN τ ), one for each τ -type, has been developed based on discriminating variables discussed in Ref. [10].These variables exploit differences between hadronically decaying τ lep-tons and jets resulting from the fragmentation of quarks and gluons, in particular the longitudinal and transverse shower shapes as well as isolation in the calorimeter and in the tracker.This technique has been used to perform a measurement of σ(pp → Z + X)× BR(Z → τ + τ − ) [11].
Jets are reconstructed using a fixed cone algorithm with radius R = 0.5 [12].A jet energy scale calibration obtained from transverse momentum balance in γ+jet events is applied to all jets.E T is defined as equal in magnitude and opposite in direction to the vector sum of all significant transverse energies in calorimeter cells.It is further corrected by the transverse momentum of all reconstructed muons, as well as by the energy calibration corrections applied to the transverse momenta of electrons, τ leptons and jets.A more detailed description of object reconstruction can be found in Ref. [4].
Jets from b quarks are identified using a neural network b jet tagging algorithm [13].It combines several variables that characterize the presence and properties of the secondary vertices and the tracks of high impact parameter within jets.We obtain a 54% average tagging efficiency in data for b jets containing at least two tracks with SMT hits [13], which corresponds to a 1% mistagging of jets from light quark flavors (u, d or s quarks) as b jets.The identification of b jets is only used in the ℓτ channel.
In the ℓℓ channels, the main source of background is the production of electroweak bosons that decay to charged leptons.It arises from Z/γ * → ℓ + ℓ − and Z/γ * → τ + τ − , followed by τ → ℓ ± ν ℓ ν τ with ℓ ± = e ± or µ ± , along with diboson production (W W , W Z and ZZ), when the boson decays lead to at least two charged leptons in the final state.In the ℓτ channel, the dominant background emerges from jets mimicking electrons and τ leptons, muons from semileptonic b quark decay or pion or kaon decay, and large misreconstructed E T , mainly in W +jets and multijet production.
The event selection for each channel is optimized through a minimization of the expected statistical uncertainty on the cross section using Monte Carlo (MC).Signal t t events are required to have one isolated electron or muon for the ℓτ channel or two isolated oppositely charged leptons for the ℓℓ channels.At least one jet is required to have p T > 30 GeV.All channels, except for eµ, which has the best signal over background ratio, require another jet with p T > 20 GeV.Jets are accepted in the region |η| < 2.5.Leptons are required to have p T > 15 GeV in the ℓℓ channels.A muon in the µτ channel is required to have p T > 20 GeV and an electron in the eτ channel p T > 15 GeV.Tau leptons are required to have E T > 10, 5, or 10 GeV for τ −type 1, 2 or 3 respectively.Muons are accepted in the region |η| < 2.0, while electrons must be within |η| < 1.1 or 1.5 < |η| < 2.5.In the ℓτ channels, events containing any additional isolated electron or muon passing the selection criteria used in the ℓℓ channel are rejected in order to reduce Z/γ * → ℓ + ℓ − background and to ensure that the ℓτ channels have no overlap with the ℓℓ channels.Furthermore, if more than one τ lepton is found in an event, only the one with highest τ probability (highest NN τ [10] value) is kept for further analysis.
The selection on E T is crucial for reducing the otherwise large background from Z/γ * → ℓ + ℓ − .This background is particularly important in the ee, µµ and ℓτ channels.Due to different resolutions in electron energies and muon momenta, optimization of selections leads to different criteria for the four channels.In the ee channel, events with dielectron invariant mass of M ee < 15 GeV or 84 < M ee < 100 GeV are rejected.For M ee > 100 GeV (15 < M ee < 84 GeV), they are required to have E T > 35 GeV ( E T > 45 GeV).The final selection in the eµ channel requires the scalar sum of the most energetic (leading) lepton p T and the p T of the single jet (two most energetic jets) to be H T > 105 GeV (H T > 115 GeV).This requirement rejects the largest backgrounds in this final state, which arise from Z/γ * → τ + τ − and diboson production.In the µµ channel, events are required to have E T > 40 GeV.The dimuon invariant mass M µµ must be larger than 30 GeV.To reduce Z/γ * → µ + µ − background, we define a likelihood ratio variable based on the per-event E T probability distribution, calculated from the expected resolution on E T and the energies of electrons, muons and jets.This E T likelihood ratio variable is required to be larger than 5.For ℓτ channels, events are required to have 15 < E T < 200 GeV.To reduce the multijet background, a two dimensional selection is applied in the (∆φ( E T , ℓ), E T ) plane where ∆φ( E T , ℓ) is the difference between the azimuthal angle of the E T direction and of the lepton: ∆φ( E T , e) > 2.2 − 0.045 × E T (GeV) in the eτ channel and ∆φ( E T , µ) > 2.1 − 0.035 × E T (GeV) in the µτ channel.Furthermore, in the eτ channel, events with electrons and E T collinear are rejected by requiring cos(∆φ( E T , e)) < 0.9.In the µτ channel, events with a second non-isolated muon are rejected if the invariant mass of the two muons lies in the mass range 70 < M µµ < 100 GeV.The final selection in the ℓτ channels requires at least one b-tagged jet.
The acceptance and efficiency for the t t signal are derived from a combination of MC simulation and data.Top quark pair production is simulated using the alpgen [14] matrix element generator, assuming m t =170 GeV.These events are processed through pythia [15] to simulate fragmentation, hadronization and particle decays and then passed through a geant3 [16] based simulation of the D0 detector.Data events from random pp crossings are superimposed on MC generated events to reproduce detector noise and luminosity dependent effects in data.The same reconstruction process is applied to both data and MC events to determine the selection efficiencies.Lepton trigger and identification efficiencies, as well as lepton momentum resolution, are derived from Z/γ * → ℓ + ℓ − data by strictly identifying one charged lepton as tag and using the other charged lepton as a probe.The efficiencies are studied in different detector regions and as a function of the number of jets.The lepton and jet reconstruction efficiencies, as well as the lepton, jet energy and E T resolutions in the MC are adjusted to the values measured in data.
Background contributions are also determined from a combination of MC simulation and data.The selection efficiencies for the Z/γ * and W +jets backgrounds are estimated using MC samples generated by alpgen interfaced with pythia while for diboson production they are estimated using pythia.The Z/γ * and diboson processes are generated at leading order (LO) and are normalized to the next-to-next-to-leading order (NNLO) inclusive cross section and to the next-to-leading order (NLO) inclusive cross sections, respectively [17,18].As the p T distribution of the Z boson is not well described in the alpgen simulation, the p T spectrum was reweighted to reproduce that in Z → e + e − data in the different jet multiplicities.
In the ℓτ channel, the simulated inclusive background from W + ≥ 2 jet events is normalized by fitting the transverse mass distribution [19] of the isolated lepton and E T to data.We estimate the multijet background from data using events having an electron or muon and a τ lepton of the same-charge (after subtracting contributions from W and same-charge t t MC events).The t t contributions to the same-charge sample result either from a jet reconstructed as a τ lepton or from a misidentification of the charge of the τ lepton.Contributions from Z/γ * and diboson events to the same-charge sample are negligible.
In the ℓℓ channel, the instrumental background is also determined from data.False electrons can arise from jets comprised of an energetic π 0 or η, and an overlapping track from γ → e + e − conversion.In the ee and eµ channels, the background from false electrons is fitted to the distribution of the electron likelihood discriminant in the data as done in Ref. [4].The shape of the electron likelihood is determined for true electrons in a Z/γ * → e + e − data sample.The shape of the electron likelihood for background electrons is then determined using a data sample with low E T dominated by false electrons.An isolated muon can be mimicked by a muon in a jet when the jet is not reconstructed.We measure the fraction f µ of muons that appear isolated in a sample enriched in semileptonic decays of heavy flavor quarks and in pion or kaon semileptonic in-flight decays.In this sample, one of the muons is required not to be isolated while the second serves as a probe.In the µµ channel the number of events with a false isolated muon that contribute to the final sample is evaluated as in Ref. [4].In the eµ channel, the contribution from events with a true electron and a false isolated muon is given by the number of events in a sample without a muon isolation requirement (where the electron and the muon have the same charge) multiplied by the rate f µ introduced above.Although the Z/γ * → ℓ + ℓ − processes do not lead to high p T neutrinos, they can have large E T from mismeasurements.The E T spectra from Z/γ * → ℓ + ℓ − data and the MC agree well, after jet, electron and muon resolutions are adjusted in the MC to match the resolutions observed in data.
In the ℓτ channel, instrumental background can arise from a candidate electron that does not satisfy electron selection criteria but can mimic the signatures of the type 2 τ lepton.To discriminate between the τ -type 2 leptons and electrons, we use another neural network (NN e ) [10] along with NN τ .The NN e neural network relies on a subset of the input variables to NN τ and on other variables based on the properties of the electromagnetic clusters and on the correlation between them and those of the leading track of the τ lepton.In addition, in the eτ channel, τ lepton candidates with track φ < 0.02 radian from the nearest border of the calorimeter module are removed since they are more likely to come from misreconstructed electrons.A τ lepton can also be mimicked by a jet.The corresponding rate for such misidentification is determined through a correction factor from a comparison of W +jets MC samples to e+jets data, where the estimated contribution from multijet events as well as from Z → e + e − , Z → τ + τ − and t t have been subtracted via MC.This correction factor is then applied to the W +jets and t t → ℓ+jets samples.
The expected number of background and signal events and the number observed in data as well as the selection efficiencies and luminosities are summarized for all channels in Table I. Figure 1 shows the expected and observed distributions for several observables in the combined ℓℓ and ℓτ channels.Figure 2 shows distributions in τ -types and E T of the τ lepton in the ℓτ channels.The systematic uncertainty on the measured t t pro-TABLE I: Expected number of background and signal events, observed number of events in data, selection efficiencies and luminosities for all dilepton channels.Uncertainties include both statistical and systematic contributions (excluding luminosity uncertainty of 6.1% [20]).The signal efficiency is quoted for mt=170 GeV and the expected number of signal events for σ t t = 7.9 pb [21].duction cross section in the dilepton channel is obtained by varying the efficiencies and background contributions within their uncertainties, taking all correlations among the different channels and background contributions into account.The statistical uncertainties on MC and backgrounds are treated as uncorrelated among channels, while other sources of systematic uncertainty are treated as correlated.The dominant systematic uncertainties are summarized in Table II for individual channels and in Table III for the combination of channels.

Number of Jets
The systematic uncertainties on trigger efficiencies (∼ 2% of the cross section) are derived from data.Various sources of bias are investigated, and the resulting changes in trigger efficiencies are included as systematic uncertainties.
The systematic uncertainty for identifying τ lepton (∼ 5.5% of the ℓτ cross section) arises dominantly from the uncertainty on the data to MC agreement and from the statistical uncertainty on the correction factor for jets mimicking τ leptons.The systematic uncertainty for the τ lepton energy scale (∼ 6% of the ℓτ cross section) is estimated from the calorimeter's response to single pi-ons [10].
The systematic uncertainties from the reconstruction and resolution of jets (∼ 1%) are determined from the uncertainty on the data/MC correction factors.The uncertainty on the calibration of jet energy (∼ 4%) is propagated to the predicted background and to the efficiency for t t signal.
The uncertainty on b tagging specific to the ℓτ channel (∼ 4.5%) is evaluated by shifting the jet tagging probability within its uncertainty.The flavor dependent uncertainties are evaluated by changing the parametrization of the tagging probability for different types of jets (b, c and light jets).
The uncertainty on theoretical modeling of t t production (∼ 5%) is estimated by comparing the acceptance of the two MC programs, pythia and alpgen.The full difference in the final result is quoted as the systematic uncertainty.Half of the difference between unity and the ratio of the NLO diboson cross section to the LO diboson cross section (used to scale the diboson cross sections in pythia) is taken as a systematic uncertainty for the diboson background.The systematic uncertainty on the normalization of the Z/γ * background is estimated by propagating the uncertainty on the p T reweighting function of the Z boson.
The systematic uncertainties on electron background in the eµ and ee channels are evaluated using the shape dependence of the electron likelihood discriminant on electron p T and the detector occupancy (number of jets).
Other smaller sources of systematic uncertainties (∼ 2.5%) arise from vertex identification, parton distribution functions and E T modeling.The luminosity uncertainty (∼ 6%) [20] on the cross sections is evaluated taking into account both the uncertainty on the predicted number of signal and background events.
Cross sections for individual channels are extracted using a likelihood technique described in Ref. [4].The results are presented in Table IV.All cross sections agree within their uncertainties.The combined result is ob-  tained by minimizing the sum of negative log-likelihood functions from the five channels.All systematic uncertainties are incorporated in the fit as "nuisance parameters" [22] that can affect the central value of the cross section.The result from combining the ee, eµ and µµ (ℓℓ) channels is: Both results are derived for m t =170 GeV.These represent the most precise t t cross section measurements published so far in the dilepton channel.
To improve the statistical uncertainty in the ℓτ channels, the signal acceptance for all the ℓτ results quoted above includes contributions from t t events in which the τ selection is satisfied by jets mimicked τ leptons.If we now use only t t events that decay specifically to ℓτ final states, we measure: σ t t = 7.6 +4.9 −4.3 (stat) +3.5 −3.4 (syst) +1.4 −0.9 (lumi) pb.A measurement of the cross section multiplied by the branching ratio (σ t t × B) has also been performed in the ℓτ channel using the acceptance from t t events that decay specifically to ℓτ final states (where only t t events which contain a hadronically decaying τ lepton at generator level are considered).The expected contribution from other t t events is normalized using the theoretical cross section [21].In the combined eτ and µτ channels we obtain the value for σ t t × B(t t → ℓτ b b): σ t t × B = 0.13 +0.09 −0.08 (stat) +0.06 −0.06 (syst) +0.02 −0.02 (lumi) pb, for m t =170 GeV, which is in good agreement with the SM expectation of 0.14 ± 0.02 pb [3,23].Dividing the σ t t × B(t t → ℓτ b b) measurement by the SM expectation, we can set an upper limit on the ratio of 2.3 at 95% confidence level (CL).
The value of quark masses depends on the perturbative QCD renormalization scheme, and can differ considerably for, e.g., pole mass or MS mass definitions [24].It is therefore important to extract the mass of the top quark through a well-defined renormalization scheme.Direct top quark mass measurements compare measured distributions to distributions simulated by LO MC generators.Like any LO calculation, these MC generators are not precise enough to fix the renormalization scheme, which leads to uncertainty in the input mass definition.In the present analysis, we extract the mass of the top quark using the measured top pair production cross section.This has the advantage of not relying on simulation of the t t signal, except for determining detection efficiency.The sensitivity to any differences between the pole mass and the mass used in the MC simulation is thereby reduced relative to a direct mass measurement.We compare our result to fully inclusive t t cross sections calculated in higher-order QCD that includes soft gluon resummations, which are currently the most complete calculations available.The cross sections are computed using the pole mass definition for the top quark which is thus the parameter extracted here.
We extract the t t cross section σ t t combining the ℓℓ and ℓτ channels using the selections described above and different values of the top quark mass for calculating detection efficiencies in fully simulated t t events.The result is extracted using the same function as given in Ref. [23], Figure 3 compares this parameterization of the combined measurement with a prediction including soft gluon resummation effects [23] and an approximate NNLO computation [25].For the theoretical computation we plot a 68% CL interval that we determine based on Ref. [23] or [25].The uncertainty from the ambiguity in the scale of QCD (which are varied from m t /2 to 2m t ) is represented by a likelihood function that is constant within the ranges given in Ref. [23] or [25] and vanishes elsewhere.The uncertainty due to the parton distribution functions is represented by a Gaussian likelihood, with rms equal to the uncertainty determined in Ref. [23] or [25].For every value of the mass of the top quark, we form a joint normalized likelihood function based on the theoretical likelihoods and on a likelihood for the measurement constructed from a Gaussian with rms equal to the total experimental uncertainty [26].We find m t = 171.5 +9.9 −8.8 GeV at 68% CL using Ref. [23] and m t = 173.3+9.8  −8.6 GeV at 68% CL using Ref. [25].These values are in agreement with the current world average of m t = 172.4± 1.2 GeV [27], indicating that any deviation of the directly measured mass from the true pole mass of the mass of the top quark is ∼ < 10 GeV at 68% CL.
In summary, we described in this Letter the measurement of the t t cross section in the dilepton and lepton+τ channels using approximately 1 fb −1 of D0 data.The combined cross section is measured to be: σ t t = 7.5 +1.0 −1.0 (stat) +0.7 −0.6 (syst) +0.6 −0.5 (lumi) pb for a mass of the top quark of m t =170 GeV, in agreement with the QCD prediction.We measured σ t t × B(t t → ℓτ b b) = 0.13 +0.09 −0.08 (stat) +0.06 −0.06 (syst) +0.02 −0.02 (lumi) pb which agrees with the SM expectation.Using both the t t cross section measurement and the theoretical prediction, we extract the mass of the top quark: m t = 173.3+9.8  −8.6 GeV which is consistent with the mass of the top quark from direct measurements.

FIG. 1 :
FIG.1:Expected and observed distributions for the combined ℓℓ and ℓτ channels for events with ≥ 1 jet (eµ) or ≥ 2 jets (ee, µµ, ℓτ ) following all selections for (a) the number of jets per event, (b) leading lepton pT , (c) jet pT , and (d) ET .The t t contribution is normalized to the cross section measured in this analysis.

FIG. 2 :
FIG. 2: Expected and observed distributions in the ℓτ channel for (a) the τ -type and (b) ET of the τ lepton.The t t contribution is normalized to the cross section measured in the ℓτ channel.

TABLE II :
Summary of the effects of individual systematic uncertainties on the measured cross section (in pb).

TABLE III :
Summary of the effects of individual systematic uncertainties on the combined cross section (in pb).

TABLE IV
[23,25]Dependence of the experimental and theoretical[23,25]t t cross section on mt.The point shows the combination of the ℓℓ and ℓτ measurements presented in this Letter.