Measurement of the ratio of top-quark branching fractions top quark decaying to W boson and bottom quark and top quark decaying to W boson and any quark in the lepton+jets and dilepton channels at the Collider Detector at Fermilab

According to the Standard Model, the top quark decays to a W boson and a b quark virtually 100% of the time. The measurements of tt¯ production cross section depend strongly on that assumption. We test this hypothesis with a measurement of R = Bt→Wb/ Bt→Wq , where q is any quark, using a combination of event kinematics and b-tagging techniques. The measurement is carried out using a data, sample produced in pp¯ collisions at 1.96 TeV and collected at the Collider Detector at Fermilab between March 2002 and September 2003 with an integrated luminosity of ∼ 162 pb -1.; The branching ratio R is determined from the relative tt¯ tagging rates making the measurement independent of any assumption on the tt¯ cross section. Any two tagging rates are sufficient to determine R but the problem is overconstrained if more than two tagged subsamples are used. The tt¯ events are classified by the number of leptons in the final state. In lepton-plus-jets channel only one of the W bosons decays leptonically, whereas in dilepton channel both W bosons decay leptonically. The measurement of R is performed in both lepton-plus-jets and dilepton samples.; In the lepton-plus-jets channel the background is estimated using the artificial neural network (ANN) technique. The ANN approach allows us to measure the signal fraction in samples with any number of tags. By applying this method alone the branching ratio was measured to be R = 1.06+0.27-0.24stat. +/-0.16syst. .; Alternatively, the tagged background contamination in lepton-plus-jets channel is determined from a traditional a priori method using data driven and Monte Carlo based techniques. A similar approach is used to determine the tt¯ content in the dilepton sample. The combination of ANN background measurement in lepton-plusjets data sample with the a priori lepton-plus-jets and dilepton estimations leads to improved sensitivity in the final value of R = 1.12+0.21-0.19stat. +0.17-0. 13syst.; Finally, we construct the confidence level bands based on the method proposed by Feldman and Cousins. From these bands and our final measurement of R we set the lower limit of R > 0.61 at 95% confidence level.