There is strong motivation to study standard model physics using the highest-energy data provided by the Large Hadron Collider. This is aided by the process of defining clusters of hadrons to form ‘jets’. Existing jet-finders are dependent on pre-defined parameters which, to some extent, influence their properties. This thesis introduces a novel algorithm which aims to reconstruct partons outgoing from hard interactions, prior to any splitting, by concentrating solely on the highest momentum transfer scale. In this way parton properties such as fragmentation and structure functions from hadron colliders may be compared directly with results from DIS and e+e− annihilation. This original, standalone tool is named ‘traps’ - the Topological Reconstruction Algorithm for Parton Scatters. The algorithm was developed using Pythia Monte Carlo QCD events, under a pragmatic approach that assumes the model provides a good approximation to reality at both hadronic and partonic level. Various tests were made to gauge the performance of the algorithm against standard jet-finders. The infrared safety and algorithm speed were also assessed. The objective of traps is to have low sensitivity to parameters, and to be fast and robust. A high event acceptance is necessary, as maximum statistics are required where cross-sections are at their lowest. A chapter of this thesis is dedicated to a description of the author’s studies in calibration and monitoring of the timing of the ATLAS Level-1 Calorimeter Trigger system. Pulses from triggered energy are sent via largely η× φ = 0.1 × 0.1 granularity ‘trigger towers’. Synchronous triggering with 1-2 ns precision is required for the system to make an accurate energy estimate.

The low mass Drell-Yan di-muon process is investigated with the ATLAS detector at the LHC, in order to provide information that advances our knowledge of the Parton Density Functions in a region of phase space unaccessible at previous experiments. A cut-based selection of di-muon events is performed, using 2010 data with a centre of mass energy of the proton-proton collisions of 7TeV, and an integrated luminosity of 36 pb−1. The analysis probes the region of low muon transverse momentum (pT > 6GeV), and low di-muon mass region 12 < M < 66GeV. The main challenges of the analysis are the understanding of the muon isolation and the trigger efficiency. In order to reject the large QCD background the analysis relies on stringent isolation criteria. The efficiency of the chosen selection is presented in detail. The second main part of the analysis is the measurement of the trigger efficiency for low pT threshold muon triggers. This is an important aspect of the cross section measurement, since the pT spectrum of leptons from the low mass Drell-Yan process are soft and populate the trigger threshold region. In order to measure the differential cross section in mass d dM in the fiducial region of |η| < 2.4 and pT,μ1 > 9GeV and pT,μ2 > 6GeV (asymmetric analysis) or pT,μ1 > 6GeV and pT,μ2 > 6GeV (symmetric analysis), a one dimensional bin-by-bin unfolding is adopted to account for detector reconstruction and resolution effects; all the associated uncertainties are also presented. The fiducial and extrapolated differential cross sections are measured at the Born level. Dressed level corrections are also given. The measured cross sections are shown to agree with theoretical predictions within the margin of error. A precision of 9.7% is achieved in the asymmetric analysis in the lowest invariant mass bin between 12 and 17GeV; the statistical and systematic uncertainties in the same bin are 4.2% and 8.7% respectively. In the remaining mass region the total uncertainty is smaller. The luminosity error during the 2010 data taking period is estimated to be 3.4%. In addition to the Drell-Yan cross section measurement, the thesis describes the study performed in order to extract the Lorentz angle value in the ATLAS Semiconductor Tracker. The Lorentz angle is computed through the study of the SCT cluster width from both cosmic and collision data and comparison with simulation is shown. The track selection on collision data is defined and the fitting range is optimised to give robust results. Throughout this thesis the convention c = 1 is adopted.

No description available.