Etiquetage des quarks $b$ par un détecteur de vertex à pixels dans l'expérience ATLAS auprès du LHC

The capability of the \atlas~detector to tag $b$-jets is studied, usin g the impact parameter of charged tracks. High $b$-tagging performance is needed at \lhc, especially during the first years of running, in order to see evidence of the Higgs boson if its mass lies between 80 and 120 GeV$/c^{2}$. A pattern-recognition algorithm has been developed for this purpose, using a detailed simulation of the \atlas~inner detector. Track-findin g starts from the pixel detector layers. A "hyper-plane" concept allows the use of a simple tracking algorithm though the complex geome try. High track-finding efficiency and reconstruction quality ensure the discrimination of $b$-jets from other kinds of jets. After full simulation and reconstruction of $H \rightarrow b\bar{b}$, $H \rightarrow gg$, $H\rightarrow u\bar{u}$, $H \rightarrow s\bar{s}$ and $H\rightarrow c\bar{c}$ events ($m_{H}=100$ GeV$/c^{2}$), the mean rejections achieved against non-$b$-jets for a 50\% $b$-jet tagging efficiency are as follows: \vspace{-5mm} \begin{Tabhere}\centering \begin{tabular}{cccc} \hspace{3cm} & \hspace{3cm} & \hspace{3cm} & \hspace{3cm} \\ $R_{g}=39 \pm 5$ & $R_{u}=60 \pm 9$ & $R_{s}=38 \pm 5$ & $R_{c}=9 \pm 1$ \\ \end{tabular} \end{Tabhere} The analysis of data from the first radiation-hard pixel detector prototypes justifies the potential of these detectors for track-finding and high-precision impact parameter measurement at~\lhc.