The design and production of the LHCb VELO high voltage system and analysis of the Bd ⇒ K*μ+μ- rare decay

LHCb is the dedicated flavour physics experiment of the LHC. The experiment is designed for probing new physics through measurements of CP violation and rare decays. This thesis includes simulation studies of the Bd ⇒ K*μ+μ- decay. The LHCb vertex locator (VELO) is the highest precision tracking detector at the LHC and is used to identify primary and secondary vertices for the identification of the $b$ and $c$ hadrons. The VELO modules contain silicon strip detectors which must be operated under reverse bias voltage. This thesis presents the work performed on the design, production and characterisation of the VELO high voltage system. The VELO operates only 8\mm~from the LHC beam in a high radiation environment. A future upgrade will require operation at up to 10$^{16}$ n_{eq}cm^{-2}.This thesis presents a characterisation of p-type silicon sensors before and after heavy irradiations. The design of the HV system and the substantial programme of quality assurance tests performed on both the hardware and software parts of the VELO high voltage system are described. The tests cover normal operation and consideration of a range of failure scenarios. The hardware and software limits were tested. The stability of the output over time and the noise of the system were assessed. The performance is found to meet the specification, although problems at low voltage and low current operation are seen. An analysis of the current-voltage data during module production, and commissioning up to first LHC operation is given. No obvious signs of sensor degradation are seen. The VELO high voltage system complies with the safety and performance requirements in the environment where it is used and has been successfully operated throughout the first period of the LHC operations. With its current design, LHCb expects to collect approximately 10 fb^{-1} of data, running beyond this will require an upgrade of LHCb requiring more radiation hard silicon strip detectors. P-type silicon strip detectors are one possible candidate for the upgraded LHC. Tests performed on p-type detectors with four types of isolation techniques are detailed. The breakdown voltages and the full depletion voltage before irradiation are measured. Breakdown voltages above 1000 V are found from each type of isolation technique, except for an isolation scheme with individual p-stops. The average depletion voltage is approximately 170 V. The current-voltage characteristics, breakdown voltage and the charge collection measurements of five irradiated p-type detectors are measured. Approximately 30 % of the maximum charge is collected at a fluence of 10$^{16}$n_{eq}cm^{-2} for a bias voltage of 1000 V. At a fluence of 2*10$^{15}$n_{eq}cm^{-2}, the detector with p-spray could be biased at a higher voltage before breakdown than the detector with common p-stops. This follows the expectation that the p-spray technology gives better results under irradiation. The n-on-p detectors are found to be promising candidates at the fluences expected at the high luminosity upgrade of the LHC (SLHC), albeit that at the highest fluences the charge collection efficiency is significantly reduced and they must be operated at high voltages and low temperatures. The Bd ⇒ K*μ+μ- decay is a rare Flavour Changing Neutral Current decay which proceeds via the b ⇒s transition. This decay is one of the golden modes in LHCb due to its sensitivity to New Physics contributions beyond the Standard Model of particle physics from the measurement of observables, such as the forward backward asymmetry ($A_{FB}$) and its zero crossing point ($S_{0}$). The Bd ⇒ K*μ+μ- event selection is described and is used to evaluate the signal and background yields. The estimated signal yield from the simulation is 4360$^{+1160}_{-1040}$ events/2 fb^{-1}. The background rate is estimated to be 5300 $\pm$ 1800 events/2 fb^{-1}. A binned and unbinned method of extraction of $A_{FB}$ and $S_{0}$ are discussed. The unbinned method gives direct access to the value of $S_{0}$, while the binned method may introduce a small bias to the mean value of $S_{0}$ due to assumptions of fitting the data close to the crossing point. It is estimated that the $S_{0}$ can be obtained with an accuracy of $\pm$ 1.1 GeV^{2}/C^{4}, $\pm$ 0.38 GeV^{2}/C^{4} and $\pm$ 0.17 GeV^{2}/C^{4} with data samples of 0.2 fb^{-1}, 2 fb^{-1} and 10 fb^{-1}, respectively. The effect of the VELO (and other tracking detectors) misalignments on the analysis is also studied. It is found that significant misalignments can have a large effect on event selection efficiency. However, at the current level of alignment obtained from the first LHC data the effect is already expected to give a less than 10 % change. A method to study the effects of misalignments directly on $A_{FB}$ is also demonstrated