Studies of B$_{(s)}^{0}$ -> $\mu$$^{+}$$\mu$$^{-}$ with CMS experiment at LHC

The fundamental constituents of matter and the laws of physics that govern the universe are well explained by the theory known as the Standard Model (SM) of particle physics. However, the model leaves several open questions, indicating the presence of new physics. The search for new physics is one of the primary goals for the scientists behind the building of the large high-energy collider experiments. The recent trend in particle physics follows either direct or indirect search methods for the search beyond Standard Model (BSM) physics. The direct search process is very limited because an enormous amount of center-of-mass energy is required to produce a new heavy particle. In contrast, an indirect search can be done by precisely measuring the properties of the SM decay process. Two such SM processes are the decay of the B-meson (Bs0 and B0) to the dimuon final state. The leptonic decays Bs0 → μ+μ− and B0 → μ+μ− are the effective flavor-changing neutral current processes, which are disallowed at the tree level in the SM. But they only proceed through the higher-level box and penguin diagrams. Additionally, the decays are helicity suppressed, making such decays very rare compared to others. Several BSM theories predict the en- enhancement of the branching fractions and effective lifetime. So the precise measurement of the properties will constrain the parameter space of several BSM theories. In this thesis, the effort made by the CMS experiment to measure the Bs0 → μ+μ− properties and search for B0 → μ+μ− decay using the data collected by the CMS experiment in 2011 (7 TeV), 2012 (8 TeV), and 2016 (13 TeV) is presented. The signal Bs0 and B0 candidates are reconstructed using two oppositely charged muons from a displaced decay vertex. In data, along with the signal, several background sources from b-hadrons contribute, and they are categorized into peaking, semileptonic, and combinatorial backgrounds. To control the challenging peaking background coming from the misidentified hadrons (like pion, kaon, and proton), a muon identification algorithm is developed using the Boosted Decision Tree (BDT). To suppress the dominant combinatorial background, a separate BDT discriminator is developed. The branching fraction for the Bs0 → μ+μ− and B0 → μ+μ− decays are extracted by performing an unbinned maximum likelihood (UML) fit to mass, mass uncertainty, and a binary configuration to distinguish the decay signature. Similarly, the effective lifetime of the Bs0 meson in the Bs0 → μ+μ− decay is measured by performing a UML fit to mass and decay time, where decay time uncertainty is treated as a conditional observable. The fit models are validated on the pseudo-experiments to check the model-induced bias on the parameter of interest. Several systematic uncertainties are evaluated using the normalization channel (B+ → J/ψK+) and control channel (Bs0 → J/ψφ). The Bs0 → μ+μ− branching fraction measured from the simultaneous fit to multiple BDT cat- categories is (2.9 ± 0.7) × 10−9, where the uncertainty is the combination of statistical and systematic uncertainties. This is the first observation of the Bs0 → μ+μ− with 5.6 standard deviations by the CMS experiment. There is no evidence of the B0 → μ+μ− decay. There- fore, an upper limit on the branching fraction is assigned, B(B0 → μ+μ−) < 3.6 × 10−10, at 95% confidence level (CL). The Bs0 meson effective lifetime in the Bs0 → μ+μ− de- cay mode, for the first time by the CMS experiment, is τ(B0 → μ+μ−) = 1.70+0.61 ps, s −0.44 where the uncertainty includes both statistical and systematic contributions. The measured branching fraction and the effective lifetime values are consistent with the SM predictions and the other experimental results. The second part of the thesis discusses the same measurement performed using the data collected in 2016-2018 at center-of-mass energy of 13 TeV. This analysis aims to improve the precision of the branching fraction and the effective lifetime, using the previous analysis as a baseline, with more data and novel techniques. To improve the sensitivity of the analysis, the previously developed muon BDT identification algorithm is revisited, and a loose working point is selected. Secondly, the preselection cuts on the kinematic and topological variables have been loosened, and a new advanced multivariate algorithm (MVA) using the XGBoost package is trained to suppress the dominant backgrounds. For the optimized MVA working point, the branching fraction and the effective lifetime are then extracted by performing a UML fit to mass, mass uncertainty, decay time, and decay time uncertainty. Several MC corrections are derived using the B+ → J/ψK+ decay channel and applied in the signal B0 → μ+μ− MC samples. The measured B0 → μ+μ− branching fraction (s) s is B(B0 → μ+μ−) = [3.83+0.45] × 10−9 and the effective lifetime of the B0 meson is s −0.42 s τ 0 = 1.83+0.23 ps, where the uncertainties are a combination of statistical and systematic Bs −0.20 uncertainties. No signal B0 → μ+μ− is observed, and the upper limit on the branching fraction B(B0 → μ+μ−) is set to be less than 1.9×10−10 at 95 % CL. The measured values are the most precise to date and consistent with the SM predictions and other experimental results. This novel measurement will have a significant impact on our ability to comprehend the flavour anomalies reported by other experiments and to set bounds on the BSM theory.