Semi- and Dileptonic Top Pair Decays at the ATLAS Experiment

The Large Hadron Collider, starting in 2008, will be a "top factory" as top-antitop pairs will be produced with a cross section of about 830 pb at an instantaneous luminosity of 10^33 cm^-2 s^-1 during the first year. With about 30% probability top pairs decay semileptonically into a final state with four jets, lepton (electron or muon) and respective neutrino. For another 5% of the top pair events a dileptonic decay is expected. Here the final state signature is composed of two jets, two leptons and two neutrinos. In this thesis the precision for a top pair cross section measurement at the ATLAS experiment in the semileptonic and dileptonic channels with cut based analyses, applicable to the first data, was estimated. The analysis of the semileptonic decay focused especially on the study of background from QCD events either with leptons from semileptonic hadron decays or from hadrons falsely identified as electrons by the calorimeter. For the first 10 fb^-1 and assuming a fake electron probability of 10^-3 a precision for the cross section times the branching ratio of +- 0.5(stat) +- 30.4(syst) +- 24.0(lumi) pb has been estimated, corresponding to a relative precision of 16% for the theoretically predicted cross section times branching ratio of about 240 pb. The analysis in the dileptonic channel achieves a precision of +- 0.2(stat) +- 2.5(syst) +- 2.6(lumi) pb which translates into a relative error of 10% for the cross section times branching ratio of around 38 pb. The errors for both the semileptonic and the dileptonic channel are expected to improve as progress is made on the luminosity determination and the knowledge of the backgrounds from comparisons with measured data. A measurement of the cross-section ratio between the dileptonic and semileptonic channel is sensitive to scenarios of new phenomena with competitive top quark decay modes such as decays involving a charged Higgs boson. It has been estimated that such a ratio should be measurable with a relative precision of +- 0.7%(stat) +- 7.7%(sys) +- 3.1%(lumi) during the first year of ATLAS data-taking. Even though the systematic errors partially cancel in such a ratio the total uncertainty is still around 8% as the background estimates rely on theoretical predictions. This should also improve as soon as the models can be tested against measured data.