Study of γγbb final state topologies at LHC and search for high mass resonances decaying into two Higgs bosons with the CMS detector

The discovery of the Higgs boson at the LHC completes the standard model (SM) of particle interactions. Albeit very successful, the SM does not provide answers to critical questions, such as the nature of dark matter or the hierarchy problem. For this reason, theories that predict the existence of new phenomena beyond the standard model (BSM theories) have been proposed. Many BSM theories predict the existence of new particles coupled to the Higgs boson. Therefore, the consistency of the SM and the viability of these BSM theories can be probed indirectly through precision measurements of the Higgs boson properties or directly through the search for an anomalous production of Higgs bosons. These complementary approaches share to a large extent the same set of experimental procedures to reconstruct the Higgs boson from its final state products. In this thesis, we report on the search for production of Higgs boson pairs at the LHC with the CMS detector, where the two Higgs bosons decay into a photon pair and into two b quarks. This final state combines the advantage of high signal efficiency and high signal to background ratio. It also leverages in the excellent performance of the electromagnetic calorimeter (ECAL) and of the silicon tracker of CMS. An anomalous resonant production of Higgs boson pairs would indicate of the existence of new high-mass boson. Among BSM theories, models with warped extra dimensions (WED), which address specifically the hirarchy problem of the SM, provide boson candidates such as the radion (spin 0) or the graviton (spin 2) with masses that can be probed at the LHC. Among the possible di-Higgs resonant searches (e.g. HH→4b or HH→ττbb), the γγbb decay channel is the most sensitive for resonant candidates with low mass, mX<400 GeV. The resonant di-Higgs decays allow a very good rejection of spurious events due to non-resonant production of hadronic jets and photons. Upon optimization of the diphoton invariant mass resolution, of the efficiency to photons (≈90%), and of the identification of jets coming from b-quarks (≈60%) the sensitivity achieved with 19.7 fb-1 of data collected in pp collisions at a centre-of-mass energy of 8 TeV permitted to exclude a relevant fraction of the parameter space of models that predict the existence of a radion. No significant excess of events is observed above the expected background in the search mass window of 260 and 1100 GeV, and an upper maximum limit on the cross-section of 3.30 fb is placed. The data analysis has required a detailed preparatory work to understand and optimize the photon reconstruction and selection. An account of this experimental work is presented in the report, with detailed discussions of the calibration method of the ECAL, and of the optimization of the simulation of the ECAL response. In particular, the optimization of a method exploiting the azimuthal symmetry of the energy flow in pp collisions, used in the equalization of the ECAL response across the different readout channels, is presented. The photon energy calibration and identification rely to a large extent on the Monte Carlo simulation of ECAL response, which in realty is subject to variation in time due to radiation-induced damage to the crystals and photodetectors. To account for this, response variations measured from data have been modeled and appropriately propagated to the MC simulation. The adoption of this time-dependent simulation has contributed substantially also to improve the CMS sensitivity of the analysis that enabled the standalone observation of the Higgs bosons through the H→γγ decay, and the measurement of the Higgs boson mass with a systematic uncertainty of of 0.12%. This precision makes the uncertainty of the Higgs boson mass subleading in precision tests of the SM.