Search for diboson resonances in the all jets final state with CMS at √s = 13 TeV and pixel sensors development for HL-LHC

Abstract

This thesis presents the first combined search for new massive resonances decaying to either to two vector bosons or to one Higgs and one vector boson, in the all jets final state. Additionally, it provides a characterization of the spatial resolution before and after irradiation of prototype small pitch sensors for the upgrade of the CMS pixel detector at the high-luminosity LHC. Many theoretical models addressing the shortcomings of the standard model (SM), the best and yet incomplete description of elementary particles and their interactions, predict the existence of new particles with masses at the TeV scale. This thesis presents a search for such particles, which could be produced at the LHC. The adopted approach allows probing a range of such hypothetical SM extensions, in a single search and with excellent sensitivity. The analysis targets resonances with masses between 1.3 and 6 TeV decaying to pairs of bosons (WW, WZ, ZZ, WH and ZH), where the vector bosons decay hadronically and the Higgs boson decays into a pair of bottom quarks. Multiple signal production modes are investigated and, in the WH and ZH all-jets channels, vector-boson fusion (VBF) is considered for the first time. The analysis is performed on proton-proton (pp) collision data corresponding to an integrated luminosity of 138 fb^−1 recorded with the CMS experiment at the LHC at a centre-of-mass energy of 13 TeV. Because the considered resonances have large mass, the decay products of each boson are expected to be collimated into one single large radius jet. The signal extraction and data-driven background estimation methods are based on a three-dimensional maximum likelihood fit to the distributions of the reconstructed invariant mass of the two final state jets as well as the mass of the two jets, taking advantage of the signal’s resonant behaviour in these three observables. This approach was developed originally in searches for resonances decaying to two vector bosons, and has been extended for the first time in this work to WH and ZH decay modes. The analysis makes use of novel machinelearning-based algorithms to distinguish jets initiated by W/Z bosons or containing b-quark pairs from jets produced in SM background processes. The analysis approach is flexible and sensitivity is demonstrated to 16 signal hypotheses taking into account the various production and decay modes. Moreover, since the resonances can decay into multiple combinations of boson pairs, 10 such combinations are as well considered to achieve the best sensitivity. The analysis establishes a up to an order of magnitude improvement on limits on signal production cross section with respect to previous searches. In a bulk graviton model, spin-0 radions produced through gluon-gluon fusion are excluded with masses of up to 2.7 TeV, while spin-2 gravitons are excluded for masses below 1.4 TeV. When considering production exclusively through VBF, upper limits on the production cross section are set from 3 fb for a resonance mass of 1.3 TeV to 0.1 fb at 6 TeV for the Radion→ VV. Furthermore, limits on the production cross section of VBF produced G^κ=0.5 bulk combining all hadronic WW and ZZ final states are set for the first time, from 4 fb for a resonance mass of 1.3 TeV to 0.2 fb at 6 TeV. In the context of the heavy vector triplet model, spin-1 W′ and Z′ bosons produced through quark-antiquark annihilation are excluded up to 4.8 TeV, the highest mass exclusion limit to date. In addition, for resonances produced through VBF, limits on the production cross section are set from 7–10 fb at 1.3 TeV to 0.3–0.4 fb at 6 TeV, for the first time in the WH channel, for the first time in the all-jets final states and for the first time combining VV and VH decay modes. This thesis also shows, through simulation studies, how searches for such particles will greatly benefit from the high-luminosity LHC (HL-LHC) era. A factor of 2 improvement in signal efficiency and cross section limits are expected thanks to the increased granularity of the upgraded tracker that will be used for operation at the HL-LHC. Moreover, the HL-LHC will deliver up to 4000 fb^−1 of data, enabling a test of the existence of massive resonances with cross sections ≈20 times smaller than the ones probed in this thesis. A finer segmentation of the detectors improves the spatial resolution, the precision of the measurement of a particle position. However, during operation at the HL-LHC, the number of pp interactions in the same bunch crossing will be five times higher than in the last LHC data taking period. The detectors will receive a high radiation dose, which causes a degradation of their performance. The pixel detector, the system closest to the interaction point, will face the most challenging conditions. This thesis presents a detailed study of the spatial resolution of prototype planar silicon pixel sensors for operation at the HL-LHC, comparing the performance before and after irradiation. The prototypes are characterized by a 100×25μm^2 pitch, six times smaller than the one currently in use in the CMS experiment. A dedicated setup composed of three parallel planes of sensors is used to perform precise measurements in the shorter pitch direction. The measurements were performed in the DESY II test beam facility with a 5 GeV electron beam. A sensor irradiated with neutrons to φeq = 3.6 × 10^15 cm^−2, more than 70% of the full lifetime fluence of the second barrel layer, two sensors irradiated with protons to φeq = 2.1 × 10^15 cm^−2, corresponding to the full lifetime fluence of the third layer, and several non-irradiated sensors were tested in this study. The thesis presents a review of the different quantities adopted in literature to define the spatial resolution and introduces a new variable. The measurements were repeated for different beam incidence angles to determine the angle providing the best resolution. A spatial resolution of 2.4 ± 0.1 μm (4.1 ± 0.1 μm) was found at the optimal angle for a non-irradiated (proton-irradiated) sensor when correcting for multiple scattering. The results show that the tested sensors are suitable candidates for high precision measurements at the HL-LHC.