Search for the Higgs boson at ATLAS/LHC in WH associated production and decay to b quark pairs

The Higgs mechanism was incorporated in the Standard Model of elementary particles and interactions in the 1960's to solve the existent conflict between massive particles and conservation laws of particle physics. A consequence of this mechanism is the prediction of a new fundamental particle, the Higgs boson, observed for the first time in 2012 by the ATLAS and CMS experiments at the Large Hadron Collider. This thesis describes the search for the Higgs decay into a pair of $b-$quarks with the ATLAS experiment, using $pp$ collision events with an 8~TeV center-of-mass energy provided by the LHC in 2012. Although the branching fraction of the $H\rightarrow b\bar{b}$ decay is dominant ($BR(H\rightarrow b\bar{b})=$57.7\% for $m_H=125~GeV$), this decay mode was not yet observed. The search is particularly challenging given the huge amount of background events containing jets. To reduce this background, the Higgs production associated with a $W/Z$ boson is usually explored, as the leptons resulting from the $W/Z$ decay can effectively trigger the signal. The $W$ associated production with the $W$ boson decaying leptonically is considered. The data analysis searches for events compatible with the $WH\rightarrow l\nu bb$ signal topology: one electron or muon, missing transverse energy associated with the undetected neutrino and two jets resulting from $b-$ quark fragmentation. Events containing jets and charged leptons, as top-quark production and $W+$jets, are the main backgrounds of the analysis. Since their production cross-section is much larger than the signal cross-section, the resulting signal-to-background proportion, $S/\sqrt{B}$ is only 0.3. The analysis comprehends a Multivariate technique, Boosted Decision Tree (BDT), to exploit correlations in the event observables aiming at increasing the sensitivity to the signal. A study that resulted in a 12\% gain in the BDT performance was carried on. Samples of signal and background simulated in the same conditions as data are also analysed. Given the small $S/\sqrt{B}$, it was indispensable to verify that the simulation models correctly the background processes, and to evaluate the systematical uncertainties associated with their prediction. In this context, a study to determine the systematic uncertainties of the single top background modelling was conducted. The ratio between the observed signal event rate and the Standard Model prediction was $1.65\,{^{+0.58}_{-0.56}}(\rm{stat}){^{+0.58}_{-0.48}}(\rm{syst})=1.65{^{+0.82}_{-0.74}}$, and therefore the measurement is compatible with the SM prediction within uncertainties. The signal significance, representing the compatibility between the data observation and the background-only hypothesis, corresponds to the gaussian probability of observing a value larger than 2.02 standard deviations and is not sufficient to state the observation of the $WH\rightarrow l\nu bb$ process.