Towards measurement of the W → τν cross section at √s = 7 TeV using the ATLAS detector

The discovery of the neutron and the nuclear model based on the Rutherford scattering experiment leads to the conclusion that matter is constituted with fundamental particles. It was thought initially that they were only nucleus and electrons. However, with the development of particle accelerators, the fundamental structure of the matter is being explored through higher and higher energy particle beams. Now we know that the matter we know is composed of 28 elementary particles (including particles and anti-particles) interacting with each others through four fundamental forces. On the other hand, many theoretical models were established to explain the universe according to the experimental results. The standard model (SM) is found to be the most mature theory as it explains how the elementary particles interact with each others. However, it also has predicted the Higgs boson which has yet to be discovered. The Large Hadron Collider (LHC) has been built at the European laboratory for particle physics (CERN) with the aim of searching the SM Higgs boson, as well as new particles that might manifest itself. It is designed to collide proton-proton (p-p) at a center-of-mass energy of sqrt(s) = 14 TeV, with a peak instantaneous luminosity of 1034 cm−2s−1. A Toroidal LHC ApparatuS (ATLAS) is one of the general purpose experiments designed for the LHC. It detects particles produced in p-p collisions and would be able to detect any new ones which were not found by previous experiments. It is composed of several sub-detectors including trackers used to reconstruct the tracks of charged particles, calorimeters used to measure the energies of the particles and finally muon chambers used to detect muons and measure their momentum. One of the interesting particles described by the SM is the tau (τ) lepton. Its physical properties (mass, lifetime, decays, etc.) allow it to be part of many processes within or beyond the predictions of the SM. For example, the SM Higgs boson decays to τ+τ− final states with branching ratio sig- nificant enough to make this channel primordial for a discovery within the mass range 115 < mH <140 GeV. However, the challenge at the hadron collider is the background understanding and measurement, which is the first and indispensable work to discover any new particle. Especially for the channel with the τ lepton decaying hadronically, the signal suffer from overwhelming background from QCD multi-jets which has the largest pro- duction cross section in p-p collisions at the LHC. Therefore, understanding the performance of τ lepton reconstruction and identification is essential for all the channels related to τ leptons. To understand its performance, se- lecting pure signal samples is important. One of the largest sources of τ leptons is W → τν process. So getting the W → τν signal and measuring its cross section is an essential issue. The aim of this thesis is to cover some of the topics related to the ATLAS detector performance as well as studies of τ lepton identification and its first application to the W → τν cross section measurement. After a brief introduction in chapter 1, the SM of elementary particles is described in chapter 2, as well as the theoretical motivation behind the interests in τ lepton final states physics. Chapter 3 describes the LHC and the ATLAS detector with performance of the liquid argon calorimeter (LAr) as it is an important part of the calorime- ter system. It consists of the electromagnetic calorimeter (EM), the end- caps of hadron calorimeter (HEC) and forward calorimeter (FCAL). The LAr calorimeter plays a crucial role in hadronic τ decay identification as it measures some of the physical quantities, combined with the tracking information, provide a powerful tool in separating its signal (i.e. τ lepton) from QCD jet background (see chapter 5). Several studies using the col- lected data from cosmic rays, 900 GeV and p-p 7 TeV collisions have shown good LAr calorimeter performance that is in accordance with expectations in order to achieve the optimal measurements necessary to physics studies of complicated processes such as W → τν. Chapter 4 shows the calorimeter response and energy resolution to pions using data collected during the 2004 test beam. During these tests a segment of the full ATLAS calorimeter system, as well as other sub-detectors, has been studied in order to understand its performance in a more realistic environment. First, random trigger has been used to measure the electronic noise for each cell of the calorimeter segment. The results show that the average noise is 12 MeV in the presampler, 12 MeV in the first layer, 28 MeV in the second layer and 22 MeV in the third layer of the LAr calorimeter, 30 MeV in the first layer, 30 MeV in the second layer and 25 MeV in the third layer of the tile calorimeter. Low energy (injection beam momentum < 10 GeV) and high energy ((injection beam momentum < 20 GeV) pion beams have been used to measure the LAr calorimeter response and resolution, as well as its dependance on beam energy and the position with respect to the incoming particles (η dependence). The LAr calorimeter cell-time to trigger-time (timing) has been estimated with the requirement that the cell energy is higher than 2.5 GeV and has been proved to be within 5 ns for good runs. The results have shown that the calorimeter response increases with the pion energy and is slightly dependent on the η position of the incoming pions. The energy resolution has also been studied and found to be linear with respect to 1/sqrt(Ebeam), which is in agreement with the Ebeam assumption that it is dominated by fluctuations inherent in the development of electromagnetic and hadronic showers. Chapter 5 describes the hadronically decaying τ lepton reconstruction and identification performance in the 900 GeV and 7 TeV collision data. This is achieved by comparing and optimizing the reconstruction algorithms us- ing Monte Carlo (MC) events from hadronically decaying τ lepton final states (W → τν) as well as background processes such as QCD multi-jets. Distributions of the variables used in the identification methods show good consistency between data and MC, and provide a good separation between hadronic decay τ lepton from QCD jets. Three identification methods are developed and studied with data and MC events: the cut-based method, boosted decision tree (BDT) and likelihood (LLH) method. The cut-based method gives an identification efficiencies about 70% for loose level, 50% for medium level and 30% for tight level. These three levels of identification have been defined according to acceptable background contribution for dif- ferent physics analysis. The BDT and LLH method give higher efficiencies compared to the cut-based one. Chapter 6 presents the selection of W → τ ν signal events, where the τ lepton decays haronically and the methods to estimated the cross section of this process at the LHC. This analysis shows the performance of the track-based missing transverse momentum (pTmiss) and the use of its correlation to the T calorimeter-based missing transverse energy (ETmiss) in order to reject QCD multi-jets background. For signal events, where there are neutrinos in the final states, produce a real Emiss while QCD multi-jets background events have an Emiss mainly from mis-measurements of jets energy. Therefore, independent measurements from calorimeters (ETmiss) and trackers (pTmiss) are strong correlated for the signal events. Using this property, 60% of QCD multi-jets background events can be rejected. A new quantity, called new defined missing energy significance (SMETnew) has been defined and used for suppressing QCD multi-jets background as well as estimating its contamination to the selected signal events directly from data. SMETnew is defined as ETmiss[GeV]/(0.5*sqrt(sumpT[GeV])),where sumpT is the total transverse momentum of the event measured from all reconstructed good quality tracks from the primary vertex. SMETnew has proved to be less sensitive to pile-up events when the LHC instantaneous luminosity increases and hence shows more stable measurements of the W → τν cross section. This is due to the fact that sumpT from one vertex is not effected by others vertices. In the analysis, 1545 ± 76 (stat) signal events are selected with an estimated QCD multi-jets background events of 642 ± 45 (stat). Based on these numbers, a very preliminary measurement of the cross section (W → τν → τhadν) 5.96 ± 0.26 nb is obtained with statistic uncertainty shown only.