Gaetano de Marino ; Francesco Forti ; Giulia Casarosa

2019
Università di Pisa Pisa

Abstract: Belle II is an international particle physics experiment located at the KEK Laboratory in Tsukuba, Japan, where the SuperKEKB accelerator operates. SuperKEKB is an asymmetric electron-positron B-factory, designed to operate at a center-of-mass energy of 10.58 GeV, corresponding to the mass of the Υ (4S) vector meson resonance, which decays into B meson-antimeson pairs nearly 100% of the times. The asymmetry in the beam energies results in a boost of the center of mass in the laboratory frame, necessary to study time-dependent decay rates and CP asymmetries. While the previous experiments (BABAR and Belle) at first-generation B-factories (KEKB and PEP-II) were mainly dedicated to the measurement of CP violation in the B system, Belle II will focus on precision measurements of decays of bottom and charm mesons and τ leptons and on the search for rare or forbidden processes that may provide indications of physics beyond the Standard Model. A significant Belle detector upgrade was necessary since the KEKB accelerator, already holding the world record for achieved luminosity, has gone through a substantial redesign. It will allow to obtain an even higher peak luminosity, providing the Belle II detector with a 50 times larger integrated luminosity of 50ab−1. This will enhance the probability of finding hints of new physics in the flavour sector and will surely improve the precision on measurements of the Standard Model parameters. My research project aims to measure the D0 lifetime with the first processed data of the Belle II experiment, collected during the period of data taking between March 2019 and June 2019 and corresponding to an integrated luminosity of 2.6fb-1. The target channel of the analysis is D*+→(D0→K−π+)π+s +c.c., whose cross section at √s = mΥ(4S)c^2 is ∼9 pb. The objective is to demonstrate the vertexing capability of Belle II as an essential element for time-dependent measurements and for a large fraction of the physics program. This measurement requires both high resolution on track momenta and the ability of resolving the production and decay vertices of the D0’s, which is rather challenging since the average flight distance of D0’s from D* decays is ∼220 μm. D* candidates are selected from ccbar events and not from B decays because they add a non-negligible flight length to those of the D0’s and would induce e bias in the measurement. The momentum of D*’s coming from the decay of B mesons has a maximum value that can be used as a threshold to discard the processes involving B decays. The pion from D* decay is traditionally referred to as “slow” pion since its phase space is very limited due to the small mass difference between the D* and the D0. This results in a very narrow distribution for the released energy of the decay, defined as Q =m(D0πs)−m(Kπ)− m(πs). The main elements of the analysis examined in this thesis are: the calibrations, the analysis tools and the reconstruction performance, mainly related to tracking (efficiency and impact parameter resolution), vertexing and particle identification. A precise knowledge of the position of the Interaction Point (IP) and the size of the beam spot, i.e the region where the two colliding beams overlap, is needed. My task was to provide the run-dependent beam spot parameters and contribute to make them available to the analysis tools (e.g. fitters) through the access to the database. First, the IP position was reconstructed online in the context of the Data Quality Monitoring (DQM) of the experiment. Then, the same information stored for the DQM was also used for the beam spot calibration, which was as useful to the experiment as it was to our analysis. The selection for the lifetime measurement takes place in two steps: the pre-selection, centrally made by the Belle II production to provide only the candidates passing basic requirements, and the final selection, which constitutes an important part of the work for background suppression. It also required to investigate the performance of the detector since the efficiency and purity corresponding to each of the considered variables needed to be evaluated. It is important to emphasize that the selection has been studied directly on data and not on Monte Carlo (MC) since the main goal of the thesis work was to prove the good performance of the detector. Furthermore, the agreement between data and MC is still non optimal due to the early phase of the experiment. The fits to them D0 and the Q distributions have been performed in order to define the signal region where to extract the final candidates from. In the (mD0, Q) space, a sideband region is also identified, with the purpose of estimating the level of background. Characterization of signal and background was made on MC by means of the Belle II analysis tools. MC samples used for the analysis consist of 106 signal events and the equivalent of 80fb−1 for the main processes at B-factories: BB decays, continuum processes (e+e− → qqbar, q = u,d,s,c) and e+e− → τ+τ− events (collectively called generic MC). The signal was treated separately to study the resolution function and test the probability density function (PDF) for the unbinned maximum likelihood fit to the proper time distribution.The proper time t_D0 is obtained from the reconstructed momentum p⃗ of the D0, the nominal mass m_D0 and the measured distance between the production and decay vertices. The identified background sources are ccbar processes with D0 → KπγX decays and mis-reconstructed or fake πs’s attached to true D0’s to make the D* candidates, and they are both treated as signal in the final model. Combinatorial background contamination, despite being very small due to the low statistics, is included in the final fit as a separate and fixed component extracted from the sidebands. The model has been validated on the full MC sample, obtained merging the various samples. The result obtained on the 2.6fb−1 sample of early data is τD0 =(400±6(stat)±9(syst))fs. Among the possible sources of systematic uncertainties, the largest one is given by the residual misalignment of VXD or biased beam spot position calibration. These two points can have the same effect of shortening or lengthening the measured proper time. The final value of 9 fs is obtained by summing in quadrature this term with the other evaluated contributions, related to the choices made for the selection and the PDF for the fit to the proper time. The uncertainty on the measurement is dominated by the systematic error, although the available statistics is still very low. The natural next step of the analysis would be to repeat the reconstruction once improved calibrations are available, both for alignment and for the beam spot. Further studies on MC can reveal all the possible effects that such non-optimal calibrations can induce on the measurement. In addition, the analysis can be extended to other decay channels (multi-body charged and 3-body involving neutral states) but a deeper understanding of the detector performance is indispensable.

Note: Presented on 16 10 2019
Note: MSc