Measurement of the exclusive and inclusive branching fractions of B⁰_s→D^(*)+_sD^(*)-_s decays at CDF and its implications on the decay width difference in the B⁰_s- anti B⁰_s meson system

The purpose of this thesis is threefold: Firstly, new measurements of both the exclusive and semi-inclusive partial decay widths of B⁰_s→D^(*)+_sD^(*)-_s meson decays are presented. Secondly, the feasibility of extracting the unknown polarization components in B⁰_s→D^*+_sD^*-_s by partial reconstruction of this pseudo-scalar to vector-vector decay in a Monte Carlo driven analysis scheme is studied. Finally, based on the suggestions contributed by the theory community this study discusses how a measurement of the branching fraction of semi-inclusive decays B⁰_s→D^(*)+_sD^(*)-_s can contribute to gain insight about the relative decay width difference in the B⁰_s- anti B⁰_s meson system. The measurement of the exclusive branching fractions B(B⁰_s→D⁺_sD^-_s), B(B⁰_s→D^*+_sD^-_s), B(B⁰_s→D^*+_sD^*-_s), and the semi-inclusive branching fraction B(B⁰_s→D^(*)+_sD^(*)-_s) is performed based on full reconstruction of the decay B⁰_s→D⁺_sD^-_s and partial reconstruction of B⁰_s→D^*+_sD^-_s and B⁰_s→D^*+_sD^*-_s from a hadronic data sample 6.8 fb^-1 in size. The data used in this thesis were collected by the CDF II detector at the p anti p collider Tevatron located at the Fermi National Accelerator Laboratory near Chicago, Illinois, USA. Absolute branching fractions are determined by measuring ratios of branching fractions using B⁰_d→D⁺D^-_s as normalization channel: f_Ds^(*)Ds^(*)=(f_s)/(f_d)(B(B⁰_s→D^(*)+_sD^(*)-_s))/(B (B⁰_d→D⁺D^-_s)) The branching fraction ratio is multiplied by the ratio of the quark fragmentation fractions f_s/f_d to account for the different probabilities of B⁰_s and B⁰_s meson production at the Tevatron p anti p collision energy. The analysis described in this thesis contains several improvements over previous measurements: in the estimation of efficiencies for the reconstruction of the intermediate D⁺_s meson from two narrow mass bands of K⁺K^-π⁺ phase space for the first time the full underlying Dalitz structure of D⁺_s→K⁺K^-π⁺ is accounted for. Furthermore, for efficiently separating signal from the vast amount of combinatorial background a neural network based multivariate approach is used. Observables are directly extracted from all studied decay channels simultaneously by using a maximum likelihood based parameter estimation model that is optimized towards describing the complicated multi-component structure of the selected invariant mass spectra in a consistent manner. This is mainly achieved by consistently sharing common parameters across the full fit function. In doing so, correlations among free fit parameters are already accounted for at the fitting stage and the effort of systematic studies is reduced considerably. The thesis outline is as follows: In the first chapter, we motivate this analysis in the light of the conceptual framework of the Kobayashi-Maskawa model, and provide an overview of relevant experimental studies. Chapters 2 and 3 give a brief technical description of the experimental apparatus the used data were collected with. After detailing pre-processing of real data samples and production of simulated data in Chapter 4, Chapter 5 specifies the selection of signal candidates based on a multivariate analysis technique. In Chapter 6 a parametric model is set up to extract the figures of interest from data. To estimate systematic uncertainties, in Chapter 7 model assumptions made are scrutinized, followed by a presentation of branching fractions results and an estimate of the observable ΔΓ_s/Γ_s in Chapter 8. Chapter 9 concludes by summarizing the achievements of this study and pointing to the limitations of the ΔΓ_s/Γ_s estimation. In this context an outlook on complementary experimental strategies that might be realized in the near future is given.