Search for Supersymmetry in single-lepton final states with the ATLAS detector and Improved background model for the search of new physics

The current understanding of nature at its elementary level is condensed in a physical theory known as the Standard Model. While it has been extraordinarily successful in describing almost all known phenomena in particle physics with astonishing precision since the 1970s, it is also commonly agreed upon that it can only be an approximation of more fundamental laws of physics, which have yet to be discovered. Supersymmetry is an attractive candidate for an extension of the Standard Model since it recovers many of its deficiencies and, maybe, constitutes an important milestone on the road to a theory of everything. There are good reasons to believe that if supersymmetry exists, it will make itself apparent at the so-called weak energy scale. With the start of the Large Hadron Collider, a new era in particle physics has therefore begun, since it will be possible to thoroughly probe this scale for the first time. The experiments at the Large Hadron collider will also search for the last big missing piece of the Standard Model, the elusive Higgs boson, and potentially unveil the cause for the associated electroweak symmetry breaking, which gives rise to the masses of the particles and may well be connected with supersymmetry. This thesis presents the author’s contribution to the search for supersymmetry in final states which contain a single high-energetic electron or muon, using data recorded by the ATLAS experiment until the end of 2010. A crucial prerequisite for the search for supersymmetry or any other new physics phenomenon is the precise simulation of known Standard Model processes. Due to shortcomings in the modeling or an incomplete understanding of detector effects, the measurements will always systematically deviate from the predictions. A novel method, which incorporates such deviations into an improved model, is presented. It combines information from the simulation and from the data, obtained by control measurements, in a statistically well-defined manner.