Constraining the proton structure at ATLAS

Particle physics is at a pivotal moment: the origin of mass and new physics scenarios beyond the Standard Model or particle physics could be unveiled in the coming year. In 2007 the most powerful particl e accelerator, the Large Hadron Coolider (LHC), will start colliding proton beams reaching the ihghest energy and luminosity ever in collider particle physics. The ATLAS detector is one of two general pu rpose detectors placed along the collider ring to fully exploit the LHC potential. The theoretical uncertainties on most of the LHC physics progream are dominated by the proton structure uncertaintiy. This thesis demonstrates that $W^{\pm}$ boson productionis an ideal process to constr ain the proton strcuture uncertainty. The rapidity distributions of electrons and positrons originating respectively from the $W^-$ and $W^+$ decays have been analysed. The results show that the current uncertainty on the gluon content of the proton can be reduced by a very significant amount if the total systematic uncertainty on the experimental rapidity deistributions can be controlled to better than 5%. The background study shows that the W boson production is a very clean signal, with a small background contamination, of the order of 1%, after the application of the W selection cuts. It has been f ound that a large systematic bias, due to the electron reconstuction and identification algorithms, leading to the W event selection, distorts the rapidity distribution of the W charge asymmetry to 20% l evel on average. The electron charge misidentification rate in ATLAS has been measured applying a data-like analysis: it results to be less that 0.5% over the whole accessible rapidity spectrum. Novel techniques have been investigated and implemented to improve the computation of high order corrections and the PDF uncertainties in Monte Carlo simulations. The tracking system is a vital component of the ATLAS detector. Part of the ATLAS tracking system, the Semiconductor Tracker (SCT) of the barrel, was assembled and tested at the University of Oxford. In three years of intense work 2112 microstrip silicon modules were mounted and tested. At the end of the assembly only 0.3% of more than 3.2 million electronic channels were found defective.