The ATLAS muon spectrometer

Abstract

The Standard Model is a well established theory for elementary particle physics that describes all known elementary particles and their interactions. Except for gravity all known forces are included: the electromagnetic, weak and strong nuclear force. ATLAS is one of the two general-purpose experiments at the LHC accelerator at CERN, which is a proton-proton collider with an unprecedented nominal center-of-mass energy of 14 TeV. One of its main goals is to study the Standard Model and look for possible new physics beyond this model. Especially searched for, will be the Higgs particle, the last not-measured particle predicted by the Standard Model. Extensions to the Standard Model will be tested, e.g. supersymmetry and theories with extra dimensions. In many of the searches muons with a high transverse momentum will be crucial due to their clean signature in the detector. In this thesis the performance of the muon spectrometer of ATLAS has been studied. Four detector technologies are deployed; the Resistive Plate Chambers (RPC) and the Thin Gap Chambers (TGC) provide the trigger in respectively the barrel and the endcap regions; the precision measurements to measure the momentum are performed by three layers of the Monitored Drift Tube (MDT) chambers and for the inner forward regions by the Cathode Strip Chambers (CSC). An eight-fold toroidal superconducting magnet system provides a magnetic field of typically 0.5 T. With these detectors the muon spectrometer will provide a momentum dependent muon trigger and measure the muon momenta with a resolution between 4% for 10 GeV < pT < 500 GeV and better than 10% for transverse momenta up to 1 TeV. For each muon traversing the muon spectrometer these detectors will produce a number of position measurements. From this set of measurements, the trajectory can be reconstructed. The reconstruction is performed in several steps, first an initial pattern recognition is performed, after which the measurements in the individual chambers are locally fitted into so-called segments. From these segments track candidates are built, which are then fitted. For the initial pattern recognition, an algorithm consisting of a family of global pattern searches based on the Hough transform has been developed. All detector technologies provide a precise measurement for two dimensions (xy or Rz), while the position in the other dimension is less precise. Therefore, the algorithm performs a search in each of the two precision planes, xy and Rz. The resulting two-dimensional patterns are then combined into a three-dimensional pattern. For the Rz-hits a transform is used that accounts for the curvature of the tracks. The algorithm applies several techniques to distinguish patterns in the high background environment of ATLAS. The algorithm assigns a weight to each measurement, depending on the likelihood of it being a noise hit. Depending on the occupancy, special weighting is applied to reduce the cpu usage. It is shown that the algorithm has an excellent efficiency over a wide range of momenta and a low cpu usage in high background. The pattern recognition algorithm is part of the highly modular and recently revised MOORE reconstruction program, which is, besides Muonboy, one of the two muon standalone reconstruction programs in ATLAS. The whole chain of MOORE modules, including the segment reconstruction, segment combining and tracking has been described in detail. Its performance is demonstrated on simulated dimuon decays of Z0-bosons and J/ψ mesons. The performance metrics include efficiency, fake rates and momentum resolution. The total track efficiency and momentum resolution are well understood and competitive to the Muonboy programme. Misidentified track rates have been discussed and are shown to be under control. For the commissioning of the ATLAS detectors, cosmic muons are deployed. To utilise the possibilities of these muons completely, the tracking needs to be changed and optimised. In particular the initial pattern recognition needs to be altered to account for non-pointing tracks. For this, dedicated Hough transforms have been developed. Furthermore, several adaptations have been made to the MOORE reconstruction programme to achieve a similar efficiency as the collision muons. On simulated cosmic muon samples it has been shown that the reconstruction is well understood for different categories of events. For each of these categories, it has been shown that MOORE has a better performance than Muonboy, especially for muons with large impact parameters with respect to the interaction point. Cosmic muons have been recorded in the cavern since 2005 and the setup has been gradually extended. In 2008 all ATLAS detectors were present in the readout. For the muon spectrometer, differences between simulated and real cosmic ray muon data have been explained, with a focus on the reconstruction strategy. It has been shown that the MDT detector performance and MOORE segment and tracking reconstruction are robust and efficient for real cosmic muon data. Furthermore, it is shown that MOORE is currently the best reconstruction programme for cosmic muon data.