From the sublime to the ridiculous: top physics and minimum bias events in the ATLAS detector at the LHC

QC Physics

Wraight, Kenneth ; Buttar, Craig M.

University:University of Glasgow

Department:School of Physics and Astronomy

关键词: Particle Physics, Top Physics, Minimum Bias, Forward-Backward Correlations, ATLAS, LHC;   

Others  :  http://theses.gla.ac.uk/2723/1/2011wraightphd.pdf

来源: University of Glasgow

【 摘 要 】

This thesis is comprised of two separate physics themes, both of which involve the ATLASdetector situated at the LHC at CERN. The first constituent is a study of the top quark signalin the fully-leptonic channel for proton-proton collisions at a centre-of-mass energy of10 TeV. Here an event counting analysis is performed based on Monte Carlo simulation.This is supplemented by a study into one of the sources of systematic error. The secondcomponent is forward-backward correlations in minimum bias events. For this, there is aMonte Carlo hadron-level comparison of the correlation for 900 GeV centre-of-mass collisions,followed by a comparison of Monte Carlo predictions to data for 900 GeV and 7 TeVcollisions.Top PhysicsA measurement of the fully-leptonic ttbar cross-section in the three decay channels ee, mumu, andemu is performed on ATLAS produced fully simulated pseudo-event data-samples. Selectionrates for signal and background events consistent with ATLAS results are found alongwith the kinematic distributions of selected events. A calculation of the non-hadronic ttbarcross-section, based on the measured cross-sections, will then return the theoretical value of217:06pb used to generate the original samples, showing the closure of the pseudo-analysisprocess.A more detailed study is made of the systematic uncertainty arising from variations in theinitial (ISR) and final (FSR) state showering models, based on the Pythia event generator. Afast simulation of the ATLAS detector is used with similar object and event selection to thefully simulated case. The effect of ISR variations on the signal is found to be negligible as itis washed out in the subsequent decays of the ttbar system. However, the effect of FSR is foundto cause 5% uncertainty in the selected signal events. In addition, in the main background ofeach of the selection channels the effect of FSR is found to produce variations of up to 30%in well populated channels. The variations in signal and background measurements will thenbe used to calculate a new estimate of the systematics on the measured ttbar cross-section foreach channel.Minimum BiasA detailed study of the forward-backward (FB) correlation and event shapes of a selectionof Pythia tunes for pp collisions with CoM = 900 GeV is performed. This includes an investigationinto the sources of particle production in generated minimum bias events as wellas the component sub-processes in generated minimum bias events. The tunes are found tobe practically degenerate (within 10 - 20% variation) for the 'standard' distributions. Theinclusion of a new observable, namely the forward-backward correlation, to the standardset is recommended. The study finds that the FB-correlation and its pT and dependentvariations are able to discern differences between the selected tunes to a greater degree thanthe usual inclusive distributions. Further, the FB-correlation is found to be sensitive to theparticle production processes within the tunes, an invaluable property for the purposes ofgenerator tuning.A measurement of the forward-backward correlation for pp collision of CoM = 900 GeVand 7 TeV at the LHC using the ATLAS detector is made. The measured correlation iscompared to the predicted correlation of several ATLAS centrally produced generator tunes.A correction procedure is developed and validated on the generator samples to correct thegenerated correlation to the hadron-level correlation. This is then applied to the measuredcorrelation and a comparison of corrected data to the hadron-level predictions of the generatedtunes made. The corrected correlations at the two collision energies are compared aswell as the calculation of a global correlation at both energies. The measured and correctedcorrelations are found to lie above the predicted distributions at both energies and across theeta-range. Further investigation of measured correlation using augmented FB-correlations isrecommended.