Search for the Standard Model Higgs Boson Produced in Association with Top Quarks in the Fully Hadronic Final State at the CMS Experiment

I present my work at the CMS experiment on a search for the standard model (SM)Higgs boson produced in association with top quarks. The search is targeted towards¯final states compatible with the H → bb decay and the fully hadronic decay channelof the t¯ pair, and uses data from proton-proton collisions at a centre-of-mass energytof 13 TeV, corresponding to an integrated luminosity of 35.9 fb−1 . This is a challengingsearch with many final state particles that cannot be uniquely identified and with largecontamination from SM background processes. It is performed for the first time at CMS√and the first time anywhere at s = 13 TeV, and contributes to the overall sensitivityof the t¯H cross section measurement, which constitutes a crucial test of the SM.tThe CMS apparatus is a multipurpose detector operating at the LHC, which is ahadron collider at CERN. The CMS detector operates a 3.8 T superconducting solenoid,and includes dedicated subsystems for charged particle tracking near the interactionpoint, measurements of electromagnetic and hadronic energy deposits, and muon tracking outside the solenoid, all of which provide nearly 4π coverage. The trigger and dataacquisition system of CMS, to which I made original contributions, efficiently reducesthe event rate from the 40 MHz collision rate to around 1 kHz for permanent storageand offline analysis. I also developed dedicated jet based triggers for the fully hadronict¯H search.tI performed all aspects of the search, making original contributions to all techniquesand measurements specific to it. A jet based quark-gluon discriminator is used in anevent-based likelihood ratio for the first time in a CMS search to differentiate betweenevents containing jets originating from light-flavour quarks and events containing jetsfrom gluons. A unique method to estimate the dominant QCD multijet backgroundfrom data is developed. Selected events with 7 or more jets and 3 or more b-tagged jetsare allocated to one of 6 categories based on jet and b-tag multiplicity, with differentlevels of signal purity.A matrix element method (MEM) is used for optimal discrimination between the t¯Htsignal and SM background processes and for the ultimate signal extraction. It assigns asignal and background probability density to each event using the full event information¯and leading order matrix amplitudes of the t¯H and t¯ + bb processes. It sums over allttcombinations of jet-quark associations to reduce the uncertainty of matching the correctpairs, and it integrates over poorly measured or missing variables. A likelihood ratio ofthese two probability densities is used to form the final MEM discriminant. This is thefirst time that the MEM has been used in a fully hadronic final state at CMS.The results are interpreted via an observed t¯H signal strength relative to the SMtcross section under the assumption of mH = 125 GeV, i.e. µ = σ/σSM . A binnedmaximum likelihood fit is performed to the MEM discriminant in all categories to extracta best-fit value of µ = 0.9±1.5. This is compatible with the SM prediction of µ = 1, andˆcorresponds to observed and expected significances of 0.6 and 0.7 standard deviations,respectively. Under the background-only hypothesis, upper exclusion limits on the signalstrength of µ < 3.8 and µ < 3.1 are observed and expected, respectively, at the 95%confidence level.