Measurement of the beauty production cross section and flavour tagging calibration with the ATLAS detector at the LHC

The beauty quark is one of the fundamental particles of the SM, and is expected to play a crucial role in the ATLAS physics program.In particular the production rate of beauty quark at LHC energies is an important test of the QCD predictions. Furthermore, the identification of jets originating from $b$-quark hadronisation ($b$-tagging) is crucial for many ATLAS physics investigations, like the search for the SM Higgs Boson in the $H\to b\bar b$ decay. The work presented in this thesis makes use of the reconstruction of two physics signals, $D^{*+}$ mesons, selected through the fully hadronic decay $D^{*+}\to D^0 \pi^+ \to K^- \pi^+ \pi^+$, and $D^{*+}\mu^-$ correlations. The $b$-hadron production cross section is measured in $pp$ collisions at $\sqrt{s}=7\TeV$, using 3.3\,$\mbox{pb}^{-1}$ of integrated luminosity, collected during the 2010 LHC run. The $b$-hadrons are selected by partially reconstructing $D^{*+}\mu^- X$ final states. Differential cross sections are measured as functions of the transverse momentum and pseudorapidity. The measured production cross section for a $b$-hadron with $p_T>9$ GeV and $|\eta|<2.5$ is $32.7 \pm 0.8 (\mbox{stat.}) ^{+4.5}_{-6.8} (\mbox{syst.}) \,\,\mu\mbox{b}$, higher than the next-to-leading-order QCD predictions but consistent within the experimental and theoretical uncertainties. This thesis presents also two innovative methods, based on the $D^{*+}\mu^-$ and $D^{*+}$ samples reconstructed inside jets, to calibrate on data the $b$-tagging efficiency respectively on $b$-jets and $c$-jets. This is a fundamental requirement to use $b$-tagging in ATLAS physics analyses. A sample of $D^{*+}\mu^-$ correlations reconstructed inside jets allows to reconstruct a very pure sample of $b$-jets, very suitable to be used to calibrate the $b$-jet tagging efficiency on data. This has been used with 2010 data, showing similar performance to the already established methods, and thus proving to be an excellent complementary technique to other standard calibration methods. No calibrations have been ever developed to measure $c$-jet tagging efficiency. During this thesis, a method suited to perform this task has been developed. It is based on the selection of a sample of $D^{*+}$ mesons reconstructed inside jets, which allows to collect a sample of $c$-jets with high purity. This method provided the first $c$-jet tagging calibration in ATLAS and its results have been used to correct the simulations of all physics analyses, implying $b$-tagging, performed with 2011 and 2012 data samples.