Detector optimization and physics performance of the CMS Phase-2 Endcap Timing Layer

The High Luminosity Large Hadron Collider (HL-LHC) is an extremely challenging upgrade of the present LHC since it will increase the instantaneous luminosity by a factor of 5. As a result, in CMS, the number of concurrent events in each bunch crossing will be 140-200, and new tools are needed to maintain at HL-LHC the same performance obtained at LHC. One of the major consequences of the high pile-up is that two events are considered as one 15% of the time since they appear to originate from the same vertex. As vertices are overlapped in space but not in time, a 30 ps track resolution will allow a reduction of vertex merging from 15% in space to 1% in space-time. For this reason, the CMS Collaboration has decided to install a Minimum Ionizing Particle (MIP) Timing Detector (MTD), providing time information to associate charged tracks with the correct interaction vertices. The MTD consists of two detectors: the Barrel Timing Layer (BTL), equipped with LYSO crystals read out by SiPM, and the Endcap Timing Layer (ETL), instrumented with Ultra Fast Silicon Detectors. The choice of different technologies for the two regions is dictated by the radiation levels and the area to be covered with sensors. The Endcap Timing Layer will have to withstand irradiation fluences up to 1.7$\times$10$^{15}$ n$_{eq}$/cm$^2$ by the end of its lifetime, and silicon sensors are the best-suited detectors to cope with such an environment. On the other hand, covering the 38 $m^2$ area of the barrel with silicon would be too expensive, and crystals represented the best option to instrument BTL. Focusing on the Endcap Timing Layer, excellent time resolution, and radiation hardness are not the only requirements that Ultra-Fast Silicon Detector should meet. The ETL sensors should also have: (i) a very high fill factor (ratio of the active and the total sensor area), (ii) an excellent gain uniformity, (iii) a uniform response with large signals until the end of HL-LHC, (iv) an occupancy per channel less than 2%, and (v) low leakage current to limit noise and power consumption. This thesis focuses on detector optimization and physics performance of the Endcap Timing Layer for the CMS Detector, including: (i) an hardware work on silicon devices to define the final sensor for CMS ETL, and (ii) a software work aimed at implementing the ETL geometry in the CMS software framework and to demonstrate the impact of MTD on CMS physics analysis. The sensors for the CMS Endcap Timing Layer The sensors chosen to instrument the CMS Endcap Timing Layer in CMS Phase-2, the Ultra-Fast Silicon Detectors, are thin silicon sensors optimized for timing based on the Low-Gain Avalanche Diode (LGAD) technology. The UFSD design provides a moderate internal gain (10-20), allowing to reach $\sim$ 30 ps timing resolution. The main goal of the experimental studies on UFSDs described in this work, from R&D to prototyping, is defining the final design for the ETL sensor. Extensive testing campaigns have been pursued to identify the optimal characteristics of the devices in terms of doping profiles and structure layout and verify that sensors produced by different vendors comply with the CMS requirements. This work includes the description of the investigations on yield and gain layer uniformity, no-gain region, timing resolution performances, and radiation resistance performed to finalize the design of the sensors for ETL. ETL geometry implementation and impact on physics analysis As the MIP Timing Layer is a completely new detector designed for CMS Phase-2, it is necessary to implement its full description inside the CMS software (CMSSW) framework. An accurate implementation of the detector geometry is fundamental to guarantee high-quality results in physics studies at HL-LHC. The geometry is presently used for simulations and will be used with HL-LHC data for reconstruction and analysis. In addition to the geometrical description of the detector, a new navigation algorithm has been developed for event reconstruction. The algorithm allows the identification of the sensor modules compatible with the position in which reconstructed tracks hit the ETL disk. On the physics analysis side, the presence of the Endcap Timing Layer, assigning timing information to the reconstructed tracks, will bring several advantages to event detection and reconstruction. The possible improvements ETL can provide in the CMS analysis at HL-LHC have been explored. In particular, the Vector Boson Scattering process with WZ final state has been chosen as a case study. Measurements of the W and Z vector bosons in longitudinally polarized states are of particular importance since they give direct access to the nature of the electroweak symmetry breaking via the exchange of a Higgs boson in the t-channel. Additionally, the WZ production can represent a probe of the non-abelian structure of the Standard Model via sensitive tests to triple and quartic gauge couplings. The investigations presented in this work focused on the impact of the new CMS Phase-2 geometry on the Quark-Gluon Likelihood calculation and the consequent discrimination between forward jets originating by quark and gluons, useful to correctly identify the signature of the VBS process with WZ final state.