000002759 001__ 2759
000002759 005__ 20211118085659.0
000002759 037__ $$aBELLE2-MTHESIS-2021-078
000002759 041__ $$aeng
000002759 100__ $$aLudovico Massaccesi
000002759 245__ $$aPerformance study of the SVD detector of Belle II and future upgrades
000002759 260__ $$aPisa$$bUniversità di Pisa$$c2021
000002759 300__ $$a97
000002759 500__ $$aPresented on 25 10 2021
000002759 502__ $$aMSc$$bPisa, Università di Pisa$$c2021
000002759 520__ $$aSuperKEKB is a B factory: an asymmetric e+e- collider designed to operate at a centre-of-mass energy of 10.58 GeV, corresponding to the peak of the Y(4S) vector meson resonance, which decays to a BBbar pair over 96% of the times. The asymmetry between the beam energies (7 GeV for e-, 4 GeV for e+) results in a boosted centre of mass in the laboratory frame, necessary to study time-dependent phenomena, such as certain CP asymmetries and decay rates. Belle II is a general-purpose particle detector located at the interaction point of SuperKEKB and managed by an international collaboration; while its predecessors (BaBar and Belle) focused mainly on the CP violation in the B system, Belle II will also concentrate on the precision measurement of the decays of bottom and charmed mesons, and tau leptons, and on the search for rare or forbidden processes that may provide evidence of effects of physics beyond the Standard Model. Belle II is expected to collect 50/ab, about 50 times more than BaBar and Belle, in several years of operation, allowing to enhance the precision of Standard Model parameters measurements, and to improve the sensitivity to possible effects from New Physics. The VerteX Detector (VXD), which is the innermost part of Belle II, is designed specifically to accurately reconstruct the four-momenta and vertices of all the charged particles in the event. It is made of two subdetectors: the PiXel Detector (PXD), with DEPFET-technology sensors in the two innermost layers, and the Silicon Vertex Detector (SVD), with double-sided silicon strip sensors in the four outermost layers. This thesis presents a study of the radiation damage in the SVD sensors, and discusses a possible upgraded detector, based on monolithic CMOS pixel sensors, for operation at high luminosity, on which physics benchmarking of some key channels has been performed. SVD silicon sensors can be damaged in different ways by the radiation field generated by SuperKEKB: the bulk displacement damage can alter the effective doping concentration, changing the depletion voltage, and can also increase the bulk-generated leakage current, which contributes to the noise; surface damage can lead to larger sensor capacitance and surface-generated leakage current, both resulting in increased noise. Algorithms to consistently monitor changes of these parameters in all of the sensors were developed; in order to study their evolution with increasing radiation damage, a method to estimate the dose in the sensors was also developed, exploiting the good correlation between SVD data and dose rates in dedicated diamond radiation sensors, which are located close to the innermost SVD sensors. SVD sensor occupancy can provide an accurate measurement of the dose by using the average deposited energy per hit (a parameter than can be easily estimated from data); however, the occupancy measurement is only possible when SVD data is being recorded, which is not always the case (for instance, data is not recorded during beam injection), leading to undetected dose. To account for the situations where SVD is exposed to radiation but not recording data, the measurements from the diamond sensors, which are always active and recording, were used: these sensor measurements show a good correlation with SVD occupancy. Several sources of bias were studied; the largest one, introduced by the trigger, was removed thanks to the introduction, in March 2021, of a new random trigger line: this resulted in an estimate smaller by a factor ~3 with respect to the one from the previous study, based on the same premises. An accurate measurement of the absorbed dose is necessary to compare the observed changes in sensor parameters to the expectations, and to extrapolate the effects of radiation damage in the future; this, in turn, is needed to plan detector operations and upgrades. An upgrade of the VXD is under study, with the goal of improving detector performance while maintaining robust operation as SuperKEKB increases its luminosity and, consequently, its backgrounds and radiation: although the presently expected background level at design luminosity is acceptable, higher safety margins would be desirable to account for the large uncertainty on these extrapolations. This thesis focuses on one of the proposals, which consists in the replacement of the whole VXD with a new vertex detector, VTX, made of five layers of Depleted Monolithic Active Pixel Sensors (DMAPS) based on a commercial CMOS technology. The all-pixel design with smaller pixels, together with the reduced material budget due to on-sensor readout electronics, should  improve the detector resolution; the smaller sensor element area and short integration time should improve background tolerance by reducing the average occupancy; the possibility to implement on-chip sparsification should reduce the necessary readout bandwidth and, together with the fact that a single type of sensor is used, simplify integration and operations. This thesis presents a benchmark study of the projected VTX performance, aimed at detector layout optimisation. A Monte Carlo simulation of "B0 -> D* mu nu -> (D0 pi) mu nu" events, with the D0 decaying either to "K pi" or to "K pi pi pi" and complete of beam-induced background, is performed both with the current VXD and with VTX; thanks to the Belle II Analysis Software Framework (BASF2), an accurate, not parametrised simulation of the events and of the detector response is possible. From the reconstruction of the simulated events, which is performed as it would be on experimental data, key detector parameters, which impact the results of physics analyses, are extracted: track finding efficiency, impact parameter resolution, event reconstruction efficiency, vertex position resolution, and flight length resolution. The comparison of VXD and VTX results shows that VTX would improve track finding efficiency by ~10%, with a large improvement for tracks with low transverse momentum (under 100 MeV/c); this results in a B0 reconstruction efficiency (for the selected channel) improved by up to ~70%. Also, the resolution of impact parameters, vertex position and D0 flight length is improved by ~20%. A good flight length resolution is a key element for analyses of time-dependent phenomena, such as neutral meson mixing and CP violation; also, the impact parameters and vertex resolution is instrumental for the rejection of background tracks. For these reasons, VTX can be expected to have a substantial positive impact on analyses results.
000002759 700__ $$aFrancesco Forti$$edir.
000002759 8560_ $$fludovico.massaccesi@pi.infn.it
000002759 8564_ $$uhttps://docs.belle2.org/record/2759/files/BELLE2-MTHESIS-2021-078.pdf
000002759 980__ $$aTHESIS