(a) Possible Tau decay diagrams.

(b) A hadronically decaying tau lepton illustration.

No distinction is made in ATLAS between a measured ‘primary’ lepton such as those produced in Z → e⁺e⁻∕μ⁺μ⁻ decay or a measured lepton produced in tau decay. The branching fractions of the leptonic and hadronic final states of tau decay have also been measured previously; the corresponding final states identified in ATLAS are shown in Table 6.1. The hadronic final states are classified into cases where either one or three reconstructed charged particle tracks are produced, since the tau lepton is electromagnetically charged with |Q| = 1. This covers about 98% of all possible hadronic tau decays.

6.3 Hadronic tau reconstruction and identification in ATLAS

Due to the Lorentz boost produced by the large mass difference between the tau lepton and its parent (typically a W or Z boson), hadronically decaying tau leptons are observed as narrow, pencil-like hadronic jets with one or three tracks, as shown in Figure 6.1(b).

6.3.1 Reconstruction

A description of τ_h reconstruction is given in Section 3.7.6.

Indentification variables

Every reconstructed hadronic jet will seed a corresponding τ_h candidate, offering no background rejection. The following variables are used to provide separation between true τ_h and hadronic jets where all energies are calibrated at the EM scale:

• Electromagnetic radius (R_EM): the transverse energy weighted shower width in the electromagnetic (EM) calorimeter:

  (6.1)

  where i is the index over EM calorimeter cells associated with the τ_h candidate (and hence its seed jet) in the pre-sampler and layers 1 and 2. The angular separation between the calorimeter cell and the τ_h axis is ΔR_i, and E_T,i^EM is the cell transverse energy.

• Hadronic shower radius (R_had): the transverse energy weighted shower width in the hadronic calorimeter:

  (6.2)

  where the index i is over the hadronic calorimeter cells and the cells in layer 3 of the EM calorimeter that are associated with the seed jet.

• Track radius (R

  track) : thepTweightedtrackwidth

  R track = ,(6.3) where the index i runs over all inner detector tracks passing the selection criteria in Section 3.7.6 within a cone of ΔR = 0.4 from the τh axis and pTi is the p T of track i. Leading track momentum fraction (ftrack): (6.4) where pT trackandpTˆτ are, respectively, the highest pT track of the τh candidate and the transverse momentum of the τh candidate. Core energy fraction f core: thefractionoftransverseenergyinaconeofradiusΔR = 0.1 from the τh candidate: (6.5) where i and j are the indices over all calorimeter cells associated with the τh candidate (and hence its seed jet) within a cone of radius ΔR = 0.1 and ΔR = 0.4 from the jet-axis, respectively. Calorimeter radius (R cal): theshowerwidthintheEMandhadroniccalorimeterweightedbytransverseenergydepositedineachcalorimeter R cal = ,(6.6) where i and j run over calorimeter cells associated with the τh candidate within a cone of radius ΔR = 0.4 of the τh axis in the EM and hadronic calorimeter, respectively. Track mass (mtracks): the invariant mass of the associated tracks. Transverse flight path significance (S Tˆ flight ): thedecaylengthsignificanceinthetransverseplaneofthesecondaryvertexforτh candidates with more than one associated track (6.7) where L Tˆ flight isthereconstructed,signeddecaylengthandΔL Tˆ flight istheestimateduncertainty. Maximum ΔR (ΔRmax): The maximal ΔR between any associated track within a cone of radius ΔR = 0.2 and the τh axis. Leading track impact parameter significance (S lead track) : S lead track = ,(6.8) where d0 is the distance of closest approach of the highest pT track to the reconstructed primary vertex in the transverse plane, and its estimated uncertainty Δd0. These variables are then used to form three different discriminants: a cut-based selection, a projective likelihood identification and a boosted decision tree (BDT) identification [58, 59, 60]. The cut values, BDT and likelihood requirements are optimised using MC signal and QCD di-jet background from data [61] at three levels of signal efficiency known as ‘loose’, ‘medium’ and ‘tight’. The values of the BDT and likelihood scores required for these points are also parametrised as a function of the τh pT, since several of the reconstruction variables’ shapes vary as a function of τh pT. 6.4 Tau mis-identification probability 6.4.1 Introduction In the search for processes such as Z∕H → ττ, several SM processes such as W → ℓν in association with one or more jets will contribute as backgrounds if a jet is mis-identified as a hadronically decaying tau lepton (diagrams for W(→ ℓν) +jets production are shown in Figure 6.4.1). Parton hadronisation and fragmentation Perturbative QCD is a theory for describing the interaction of quarks and gluons at very short distances. At long distances, the QCD coupling strength becomes much stronger and can no longer be successfully described by perturbation theory. In this region of confinement, the process of hadronisation transforms the coloured partons into colourless hadrons that are reconstructed as jets in the calorimeters. Hadronisation has yet to be understood from first principles and hence phenomenological models of a probabalistic and iterative nature are used, one example is the string fragmentation model used in the Pythia [47] MC generator. However, these models merely aim to respresent existing data and cannot claim to be correct, particularly when extrapolating jet properties to higher energy experiments. It was observed that the hadronisation models used in many generators poorly reproduce the jet shapes in LHC data and this has a dramatic effect on the tau identification algorithms. For example, the Pythia MC prediction of the yield of W+jet events in which a hadronic jet has been mis-identified as a τh is almost double that observed in data (see Chapter 7). A precise data-driven determination of the number of ‘fake’ τh candidates is therefore crucial to these channels. 6.4.2 Measuring the τh fake-rate from data The rate at which a hadronic jet is mis-identified as a hadronically decaying tau lepton is referred to as the tau mis-identification probability or tau fake-rate, fID. In this analysis, fID is defined to be: (6.9) A hadronic jet originating from a quark is more likely to be mis-identified as a τh than a hadronic jet originating from a gluon since quark-initiated jets will, on average, hadronise into a narrower η − ϕ cone and have a lower track multiplicity than jets that originate from gluons. To measure fID, events in which a photon is produced in association with a hadronic jet allow us to use the so-called tag and probe method with a large, clean sample of seed jets to be extracted from data. This is achieved by selecting a well identified photon as a ‘tag’ object that will have an associated, kinematically connected ‘probe’ hadronic jet which has not been directly subjected to any selection. The main advantage of this method is that it is largely independent of Monte Carlo or any bias introduced by using one of the tau triggers. Using the photon-jet channel also gives an estimate of the tau fake-rate appropriate for W + jet production. This is because neither the photon nor the W boson will interact via the strong force and hence the vast majority of the associated jets produced will therefore have been initiated by a quark, rather than a gluon. The leading order γ-jet production diagrams at the LHC are shown in Figure 6.4.2. An arbitrarily low fID can be achieved by sacrificing the tau identification signal efficiency, which is defined as the efficiency of identifying true τh in Monte Carlo W → τν and Z → ττ samples. Rather than attempting to minimise fID, the selection is defined for the following levels of signal efficiency: 70% signal efficiency (‘loose’ criteria); 50% signal efficiency (‘medium’ criteria); 40% signal efficiency (‘tight’ criteria). Furthermore, ‘looser’ selection corresponds to 1-prong τh candidates passing the loose criteria and 3-prong τh candidates passing the medium criteria while ‘tighter’ selection corresponds to 1-prong τh candidates passing the medium criteria and 3-prong τh candidates passing the tight criteria. 6.4.3 Event selection This analysis uses data collected during 2010, corresponding to an integrated luminosity of ∫  dt ≃ 34 pb−1 when the relevant sub-detectors were performing optimally. In order to select the γ+jet topology, events are required to pass Event-Filter level trigger in which a photon with transverse energy E T ¿ 15 GeVhasbeenidentified. Reconstructed photons in each event are required to the satisfy the following selection criteria: Cluster ET ≥ 15 GeV, |η|≤ 1.37 or 1.52 < |η| < 2.37, Well identified and isolated from other energy deposits in the calorimeter2 2This corresponds to the isEM-TightIso criteria defined in reference [62]. . Jets are selected if they have pseudorapidity |η|≤ 2.5, transverse momentum pT ≥ 15 GeV and satisfy the data quality criteria described in reference [63]. Each event is required to have exactly one photon and one jet passing these selection criteria, separated in the transverse plane such that the difference between their azimuthal angles Δϕ > π − 0.3 radians and balanced in pT such that the difference between their transverse momenta is less then half of the transverse momentum of the photon. The cut flow is summarised in Table 6.2. The identification variables for τh candidates with at least one associated inner detector track within ΔR ≤ 0.2 of a jet in events passing this selection are shown in Figure 6.4. The data / MC differences are attributed to poor hadronisation modelling in the MC, in which predictions of the hadron shower tend to be narrower than those observed in data. Cut Data 2010 Exactly one photon candidate satisfying selection 481,790 Exactly one jet satisfying selection with tau candidate satisfying selection criteria 217,101 |Δϕ(γ, jet)|≥ π − 0.3 96,882 pT-balance 90,393 (a) REM. (b) Rtrack. (c) fcore. (d) ftrack. (e) EM fraction. (f) Likelihood score. (g) BDT score. (h) Number of associated tracks. Pile-up The effects of pileup will become increasingly important as the instantaneous luminosity delivered by the LHC machine increases. To study these effects, the fake-rate was calculated for events with different amounts of pileup, by separating the events by the number of reconstructed vertices in the event. The results are shown in Figure 6.5. The fake-rate decreases with more p-p-interactions in the same bunch crossing. This result is expected, as a more crowded environment leads to a lower fake-rate due to fewer tau candidates meeting the tau identification calorimeter criteria. Due to this dependence, it was decided to calculate the fake-rate for events with different number of vertices. The bins chosen are 1-2 vertices, and ≥3 vertices. Probe jet origin After the photon-jet event selection has been applied, the probe jet sample will consist of a mixture of quark- and gluon-initiated jets. The quark fraction is defined to be the probability that a probe jet is initiated by a quark. The quark fraction is estimated using the Monte Carlo information, looking for highest pT gluon or quark within a cone of ΔR = 0.4 from the reconstructed probe jet to determine if the jet originated from a quark or a gluon. The quark fraction is shown in Figure 6.6 for γ+jet, Z+jet and QCD di-jet MC samples [51]. This procedure reproduces the cross section ratio from each matrix element to within 3%. Furthermore, good agreement is observed when calculating fID exclusively for quark- and gluon-initiated jets across the different event topologies using Monte Carlo samples [51]. This is shown in Figures 6.7(a) and 6.7(b) for quark- and gluon-initiated jets, respectively. The fake-rate for quark-initiated jets is much higher than that of gluon-initiated jets. (a) fID from quark-initiated τh candidates. (b) fID from gluon-initated τh candidates. 6.4.4 Systematic uncertainties True tau leptons in the sample The presence of events containing hadronic tau decays passing the photon-jet selection is estimated using W→ τν and Z→ ττ MC samples, requiring events to pass the selection criteria described in Section 6.4.3. Normalising the expected yield of these samples to the integrated luminosity collected leads to a negligible expected contribution due to the requirement that there be a well identified, isolated photon candidate. Di-jet contamination due to the photon selection As described in section 6.4.3, the event selection requires a well identified and isolated photon candidate. By removing the photon candidate isolation requirement, the fraction of jets initiated by quarks is expected to decrease due to the contribution from QCD di-jet events where one jet has been falsely identified as a photon. The effect on the tau fake-rate in these events and the events selected using the default photon selection is shown in Figure 6.8. As expected, the fake-rate decreases since the fraction of probe jets initiated by a quark will have decreased. A systematic uncertainty for this effect is assigned to each pT bin by taking the fID ratio with and without the photon calorimeter isolation requirement. Additional gluon-initiated probe jets As discussed in Section 6.4.3, the event selection requires that the jet and the photon are back-to-back in the transverse plane (Δϕ > π − 0.3) and balanced in pT (|ΔpT| < ). The deviation of Δϕ and ΔpT from their nominal values of π and 0, respectively, can have an effect on the fake-rate due to the increased presence of gluon probe jets i.e. events from higher order diagrams with an additional final state gluon that is selected as the probe jet. To study this effect, two sub-samples are created; one containing events which fulfil a more strict requirement on the variable in question, and one containing events which fulfil a looser requirement (up to the allowed tolerance). This is performed separately for the two variables, and the thresholds are chosen to produce sub-samples with similar numbers of events. The sub-samples are defined as follows: back-to-back cut: looser back-to-back: Δϕ < π − 0.1, tighter back-to-back: Δϕ ≥ π − 0.1. pT balance cut: loosely balanced: |ΔpT| < 0.44 ∗, tightly balanced: |ΔpT|≥ 0.44 ∗. The effect on the medium cut-based identification criteria fID can be seen in Figure 6.9. In both cases the fake-rate for the sample which fulfils the tighter requirement is slightly higher than for the sample which fulfil the looser requirement. This suggests that events in which the γ-jet pair is not well balanced in pT or are less back-to-back in the ϕ plane have a lower quark fraction. A likely explanation is that this is due to events from higher order diagrams in which the quark has radiated a gluon and one jet is outside the fiducial region of the detector (or fails the jet selection). For each of these effects, a systematic uncertainty on fID is assigned separately in each pT bin as the ratio of fID in which the photon and tau candidate pT are loosely and tightly balanced. (a) (b) 6.4.5 Results The tau fake-rate is calculated using the probe jet for each level of the tau identification algorithms listed in Section 6.4.1. Figure 6.10 shows the fake-rate for each tau identification algorithm where the hatched band represents the sum in quadrature of the systematic uncertainty due to di-jet and multijet contamination. 6.4.6 Summary A measurement of the rate at which the ATLAS τh identification algorithms mis-identify a hadronic jet as a τh has been made using data collected in 2010. The mis-identification probabilities range between 0.5% to 10% depending on the τh identification algorithm, the τh candidate pT, number of reconstructed tracks associated with the τh candidate, and the number of pileup interactions in the event. The fID measured in γ-jet data was successfully used to cross check the data-driven estimate of the W(→ ℓν) +jets background yield in the Z → ττ → ℓτh cross section measurement [64]. Chapter 7 Search for the Standard Model Higgs boson in the H →ττ and H →WW∗→ℓντν channels 7.1 Introduction This analysis is a direct search for SM Higgs boson production and subsequent decay to final states with at least one tau lepton at the ATLAS detector using data collected in 2011. In particular we search exclusively for the di-tau final state H → τ+τ− → ℓτ h 3ν. The final state with a light lepton (e,μ) and a hadronic tau decay, is expected to give the greatest sensitivity of the di-tau final states since an isolated light lepton1 1The e−− τh+, μ−− τh+ and their charge conjugate final states are considered throughout the rest of this Chapter. can be easily triggered on and the branching ratio is much higher than that to two light leptons. Another search is also made for events from the same production mechanisms and final state but through a different Higgs decay chain, in particular to a tau lepton and a light lepton via a pair of W bosons (Figure 7.1). 7.1.1 Signal and background processes Signal in H → ττ At the LHC, the SM Higgs boson is predicted to be produced mostly in the gluon-gluon fusion (gg → H) and vector-boson fusion (VBF) processes. The leading order Feynman diagrams for these processes are shown in Figures 2.4(a) and 2.4(b). In the leading-order (LO) diagrams, the two taus produced in the Higgs decay are back-to-back in the plane transverse to the beam axis. The tau decays lead to a relatively low missing transverse momentum from the undetected neutrinos. In next-to-leading-order (NLO) QCD, the Higgs boson can be produced in association with at least one hadronic jet. This has the effect of producing events with higher missing transverse momentum since the neutrinos are no longer back-to-back in the transverse plane. Signal in H → WW∗ → ℓντν In contrast, the light lepton and τh particles in signal events produced via a pair of W bosons will have a much smaller angular separation due to constraints on the final state particle helicities dictated by the spin of the Higgs boson parent, as illustrated in Figure 7.1.1. For a spin-0 parent particle such as a SM Higgs boson, the visble decay particles (in this case an ℓ − τh pair) are emitted with a small angular separation. This offers the most powerful discriminant against the dominant background processes which tend to produce these particles almost back-to-back in the plane transverse to the incoming proton beams. Background The following processes are considered as backgrounds in these analyses: Inclusive Z → ℓℓ: the largest source of background for the H → ττ analysis is irreducible, with a di-tau final state produced in Z → ττ (and to a lesser degree the Drell-Yan process qq → γ∗ → τ+τ−). The decay products in these events have similar kinematics to H → ττ signal events due to the small mass difference between the Z-boson and a light SM Higgs boson. Similar processes where the Z boson or virtual photon decay to an electron-positron (or muon-anti-muon) pair also contribute if the light lepton or any additional jet produced in the event is mis-identified as a hadronically decaying tau lepton. W(→ ℓν) +jets: the production of a W boson in association with jets forms a significant background due to the relatively large production cross section and branching ratio of a W boson decay to a charged lepton, significant missing transverse momentum from the neutrino produced in the leptonic W decay, and the fact that any jet in the event can be mis-identified as a hadronically decaying tau lepton. After making a requirement on the angular separation between the light lepton and the τh, events from W(→ ℓν) +jets are the primary background in the H → WW∗ → ℓντν analysis. tt production: the process pp → tt → W+b W−b can lead to events with missing transverse momentum, a light lepton and a tau lepton of opposite charge if both W bosons decay leptonically. There is also a small probability that a jet produced by either b quark could be mis-identified as a τh. Single top production: s or t channel single top production or single top in association with a W boson can also imitate the signal process final state if a W boson decays to a light lepton and either a W boson decays to a τh or a jet arising from a b quark is mis-identified as a τh. Diboson production: production of a pair of electroweak bosons can result in a pair of charged leptons and missing transverse momentum from the boson decays. QCD multi-jet production: events with multiple hadronic jets arising from QCD processes form an important background since the cross section is very large and there is a non-negligible probability that in a given event one jet can be mis-identified as a light lepton and another as a τh. 7.2 Datasets 7.2.1 Data The collision data were collected between March and October in 2011 during which time the LHC was operating at a centre-of-mass energy of = 7 TeV. In data periods used, all of the relevant sub-detectors were performing optimally. The dataset corresponds to an integrated luminosity of ∫  dt = 4.7 ± 0.2 fb−1 selected with high p T light-lepton triggers. 7.2.2 Simulation To understand the signal and background processes and to estimate their contributions, Monte Carlo generators were employed. For the production of W and Z bosons, the Alpgen generator [46] was used to generate matrix elements with up to five additional partons. MC@NLO [65] was used to simulate diboson, tt and single top events. The parton shower and hadronisation for these samples was performed by the Herwig [66] generator. In addition, the underlying event from additional soft QCD interactions was simulated using Jimmy [67]. The Powheg [68] generator was used in combination with Pythia [47] to generate and correctly model the gg → H and VBF production signal events with subsequent H → ττ and H → WW∗ → ℓντν decays. Events with a SM Higgs produced in association with a weak boson were simulated using Pythia [47]. For all simulated samples, TAUOLA [69] and PHOTOS [70] were used to model the tau decay and photon radiation from the charged leptons, respectively. All Monte Carlo events are fully simulated using GEANT4 [37] and reconstructed in the same way as the collision data. The cross sections for each signal and background process are summarised in Appendix A.1. 7.3 Di-tau mass reconstruction in H →ττ events Each tau decay will produce at least one undetected neutrino, making a conventional reconstruction of the original Higgs four-momentum using the decay products alone not possible. Hence, the Higgs mass cannot be fully reconstructed. In this section, three methods to reconstruct the mass of the di-tau system are explored. For the H → ττ analysis, the MMC method was used as it offers the best mass resolution of the three methods, and hence can better discriminate against the primary background process; Z → ττ. 7.3.1 Visible mass Despite the mass difference between mZ and a SM Higgs boson with 110 < mH < 150 GeV, the invariant mass distributions of the visible tau decay products, (7.1) offer little separation between Z → ττ and H → ττ due to the undetected neutrino(s) produced in each tau decay. The visible mass of the tau decay products from signal and background processes is shown in Figure 7.3.1 for events passing the selection requirements described in Section 7.5.1, for events with at least one additional hadronic jet. (a) e − τh (b) μ − τh 7.3.2 Collinear mass Since the sources of missing momentum originate from the Higgs boson, the so-called collinear mass approximation can also be used to reconstruct the Higgs mass. In the collinear mass approximation, it is assumed that the decay products of each τ are collinear with the τ in the laboratory frame since   >>  mτ, such that the taus are highly boosted in the laboratory frame. Neglecting the τ rest mass and imposing that the neutrinos in each tau decay are collinear with the visible tau-decay products, the collinear di-tau invariant mass of a system with τ → τhν and τ → ℓνν can be written (7.2) which can also be written as (7.3) where Eντh and Eνℓ is the energy (sum) of the neutrino(s) produced in the hadronic (leptonic) tau decay, χl,τh is the fraction of the tau’s momentum taken by the visible decay products and mℓτh is the invariant mass of the ℓ−τh pair. Since there are no other source of real pmiss T in the event, χℓ,τh can be also be expressed in terms of the visible tau decay products momenta and the pmiss T (7.4) and (7.5) The collinear mass of the tau decay products from signal and background processes is shown in Figure 7.3.2 for events passing the selection requirements described in Section 7.5.1 for events with at least one additional hadronic jet. The collinear mass was not used in the H → ττ analysis for two reasons. First, it tends to over-estimate the mass of the system (seen in Figure 7.3.2 as an excessive tail in the Z → ττ distribution) due to its sensitivity to the pmiss T resolution. Second, the fraction of signal events which must be discarded due to finding a non-physical mass solution is about 30%. While this is an effective cut against backgrounds, it reduces the overall sensitivity of each channel. (a) e − τh. (b) μ − τh. 7.3.3 Missing Mass Calculator Better discriminating power is obtained by reconstructing the di-tau invariant mass using the so-called Missing Mass Calculator (MMC) [71], which aims to fully reconstruct the event topology. The MMC algorithm can be thought of as an extension of the collinear mass approximation where a small opening angle is allowed between the neutrino(s) and the visible decay products of each tau decay. In this technique a set of simultaneous equations is constructed using the mass of each tau parent and orthogonal projections of the pmiss T where px missandp˙yˆmissareprojectionsofthepmiss T  vectoralongthexandyaxes,p˙vis1,2,m˙vis1,2arethemomentaandinvariantmassofthevisibletaudecayproductsandm˙τ = 1.777 GeV. The neutrino 4-momenta for each tau decay are combined into pmiss1,miss2 with invariant mass mmiss1 (in tau decays in which two neutrinos are produced i.e. leptonic tau decay), with the polar angular difference between vis1,2 and miss1,2 denoted by Δθνm1,2 for each tau decay. This system is under-constrained, with six degrees of freedom and four constraints. However, a likelihood can be obtained for a particular set of solutions using additional information the τ decay kinematics. For the MMC method, the expected three-dimensional angle between vis1,2 and miss1,2, δθ3D is used. Figure 7.5 shows the expected distribution of δθ3D for three different scenarios: hadronic tau decay with one or three associated charged hadron track(s) and leptonic tau decay. The system of equations can be solved simultaneously for any arbitrary choice of (ϕ miss1,ϕ miss2).Withthisinformation˙miss1,2andhenceΔθ3D1,2 are defined. The probability of each point in this parameter space is calculated using the δθ3D distributions shown in Figure 7.5, defined for each tau decay type and initial tau momentum from 10 < pTτ < 230 GeV in 5 GeV bins of pTτ. Since this opening angle is proportional to Lorentz boost of the τ, these distributions are parametrised as a function of the τh pT. For each region of this phase space, the δθ3D distribution is fit by a linear combination of Gaussian and Landau functions with the mean and width of pτ along with the relative Gaussian and Landau normalisations parametrised by (7.7) where ai are the parametrisation coefficients. The full PDF for (Δθ3D,pτ) can then be used to evaluate a probability for a given tau decay topology. A global event probability for a di-tau decay is then defined as (7.8) By scanning over a grid over the full range of possible (ϕ miss1,ϕ miss2),anm˙ττ distribution is obtained with each value weighted by its corresponding event. The position of the maximum of this distribution is then used to estimate the physical mττ of each event. The performance of this algorithm is highly dependent on the pmiss T resolution. Mis-measurement of the pmiss T is incorporated by convolving Pevent with Gaussian pmiss T resolution functions: (7.9) where σ is the pmiss T resolution and Δpx,y are the differences between pmiss T projections onto the x and y axis as determined from the choice of (ϕ miss1,ϕ miss2)andthatoftheevent′smeasuredpmiss T vector.Theglobaleventprobabilitythenbecomes (7.10) For a given event, the so-called ‘MMC mass’ is the invariant mass of the hadronic tau, light lepton and hypothesised neutrino four-vectors that produce the highest value of event, after a scan over mττ and the full range of (ϕ miss1,ϕ miss2). For the H → ττ analysis, the MMC mass is used since it provides the best discrimination of the signal events from the primary background, Z → ττ. In the H → WW∗ → ℓντν analysis, the visible mass is used. 7.4 Common event selection criteria 7.4.1 Trigger To preselect events from the collision data, triggers were used that select a reconstructed light lepton in a given event. For the e−τh channel, the Event Filter trigger required the event to have a cluster of transverse energy in the EM calorimeter ET > 20 GeV or ET > 22 GeV for the early or later data taking periods, respectively. This adjustment was necessary due to changing collision conditions throughout the year, with a greater number of pile-up interactions in the later periods. For the μ − τh channel, the trigger required a muon candidate with pT > 18 GeV at the Event Filter level. 7.4.2 Pileup re-weighting The total number of interactions per bunch crossing averaged over the luminosity for that block2 2Luminosity is measured in atomic units corresponding to approximately 2 minutes of ATLAS data-taking known as luminosity blocks, but the duration can vary due to run conditions and other operational issues. is defined as the average μ for that block [72]. Since object reconstruction, identification and mis-identification efficiencies are sensitive to additional interactions, it is necessary for the Monte Carlo samples to simulate additional interactions as well as the primary simulated process. The events are weighted such that the average μ distribution for each period of data-taking is replicated in the simulated datasets. 7.4.3 Muon selection A reconstructed muon candidate is required to have pT > 10 GeV with |η| < 2.5. It is further required that in a cone of radius ΔR = 0.2 around the muon the sum of additional transverse energy deposited in the electromagnetic and hadronic calorimeters be less than 4% of the candidate’s pT. Additionally, the scalar pT sum of additional inner detector tracks that have pT > 1 GeV in a cone of radius ΔR = 0.4 around the muon candidate is required to be less than 6% of the muon’s pT. Simulated data are corrected to account for observed differences in the muon pT  resolution and identification efficiency between data and simulation [73]; the corrections are ≃ 1%. If the event was triggered by a muon object, a muon candidate passing these selection criteria with pT > 20 GeV is required to be within a cone of radius ΔR = 0.2 from the trigger object. 7.4.4 Electron selection Reconstructed electrons are required to have ET > 15 GeV and |η| < 1.37 or 1.52 < |η| < 2.473 3The transition region where the barrel and endcap electromagnetic calorimeters overlap (1.37 < |η| < 1.52) is not used since the identification criteria are less effective in this region. . They are required to pass the tight selection criteria (described in Chapter 3) and are further required to be isolated in the calorimeter, such that the sum of additional transverse energy deposited in a cone of radius ΔR = 0.2 from the electron is less than 8% of the electron’s ET. Candidate electrons are also required to be isolated in the inner detector; the scalar pT sum of other inner detector tracks with pT > 1 GeV within a cone of radius ΔR = 0.4 from the electron must be less than 6% of the electron’s pT. In analogy with the muon-based corrections, simulated data are corrected to account for observed differences in the electron energy resolution and identification efficiency between data and simulation [74]. If the event was triggered by an electron object, an electron candidate passing this selection criteria with pT > 25 GeV is required to be within a cone of radius ΔR = 0.2 from the trigger object. 7.4.5 τh selection Reconstructed τh candidates are required to have pT > 20 GeV and |η| < 2.5, with exactly 1 or 3 reconstructed inner detector tracks and track charge sum equal to ±1. In the H → ττ analysis, candidates are required to pass the medium BDT-based multivariate identification requirement while τh candidates in the H → WW∗ → ℓντν analysis are required to pass the tight BDT-based multivariate identification requirements. These are described in Section 6.3 and reference [27]. 7.4.6 Hadronic jet selection Hadronic jet candidates are required to pass several data quality cuts (described in reference [63]) in order to suppress detector backgrounds and backgrounds from out-of-time pileup. Jets are also required to have pT > 25 GeV and |η| < 4.5. Finally, a cut is placed on the ‘jet vertex fraction’ (JVF), defined to be the fraction of the jet’s total track pT originating from the primary interaction vertex: the requirement is JVF > 0.75 for jets with |η| < 2.4. This final cut allows for the suppression of jet backgrounds from pile-up without raising the jet pT threshold. 7.4.7 Overlap removal Often the same final state object may be reconstructed as an object by more than one object reconstruction algorithm. In the case that different selected object candidates are reconstructed within a cone of radius ΔR = = 0.2, one of them is discarded. These overlapping candidates are resolved by selecting muons, electrons, τh and finally jets, in that order of priority. For the overlap removal procedure only, the muon calorimeter isolation requirement is ignored and the electron identification requirement is relaxed to the medium selection criteria. 7.5 Optimisation of the H →ττ event selection criteria In order to maximise the expected sensitivity of the analysis to SM Higgs boson signal events, the signal significance is maximised under variation of the selection cuts. The signal significance is defined as nS∕ where nS is the expected number of signal events and nB is the expected number of background events from all sources. The W+jets and QCD multi-jet background yields for each set of cuts is estimated using the prescriptions in Section 7.7 and Section 7.8, respectively. The varied selection criteria are: The pmiss T requirement (shown in Figure 7.5 after the event pre-selection) was varied in steps of 10 GeV from 0 to 40 GeV. (a) e − τ, 0 jet (b) μ − τ, 0 jet (c) e − τ, ≥ 1 jet (d) μ − τ, ≥ 1 jet The transverse mass of the light lepton and the pmiss T , mT , is defined to be (7.11) where pTℓ is the light lepton p T, and Δϕ[ℓ,pmiss T ] is the angular separation of the light lepton and the pmiss T   in the transverse plane. The transverse mass is effective in separating W+jets and Z+jets events from signal. An upper limit of the transverse mass of the event was varied from 20 to 40 GeV in steps of 10 GeV. Signal and background mT distributions of events with pmiss T > 20 GeV are shown in Figure 7.5. (a) e − τ, high pTmiss, 0 jet (b) μ − τ, high pTmiss, 0 jet (c) e − τ, high pTmiss, ≥ 1 jet (d) μ − τ, high pTmiss, ≥ 1 jet The degree to which the reconstructed tau candidate has been identified as a τh is varied from ‘medium’ to ‘tight’ selection. These levels correspond to a τh identification efficiency of 35% and 45%, respectively. 7.5.1 H → ττ event selection criteria After optimising for H → ττ signal, events are required to pass the following selection criteria: Exactly one identified τh and exactly one identified, isolated light lepton candidate. These candidates must have opposite charge, i.e. qτh × qℓ < 0. mT < 30 GeV. The exclusive number of additonal hadronic jets distribution of events passing these selection criteria are shown in Figure 7.5.1. (a) e − τ events. (b) μ − τ events. Additionally, Since the pT thresholds for electrons and muons is different, e − τh and μ − τh events are treated as separate signal regions, since this affects the background composition. Events with pmiss T < 20 GeV and pmiss T > 20 GeV are treated as separate signal regions. Also, events in the e−τh and μ−τh regions that do not pass the VBF region selection criteria (see Section 7.5.2) are further split by whether the number of additional jets in the event passing the selection criteria described in Section 7.4.6 is either 0 or ≥ 1. 7.5.2 VBF event selection Despite the lower cross section for the production of a Higgs boson via vector-boson-fusion, the distinct topology of these events due to the lack of colour exchange between the outgoing partons enhances discrimination of signal from background processes. In addition to the selection criteria above, the variables used to exploit these differences to create a VBF signal region using the highest and second-highest pT jets (hereafter referred to as jet 1 and jet 2, respectively) are: The di-jet invariant mass (mjet1,jet2), shown in Figure 7.9(a). Events in the VBF signal region are required to have mjet1,jet2 > 300 GeV. The pseudorapidity difference of these jets, as shown in Figure 7.9(b), is required to be larger than 3, and ηjet1 × ηjet2 < 0. A centrality requirement on the reconstructed lepton and τh pseudorapidity, such that (7.12) (a) mjet1,jet2 (b) |ηjet1 − ηjet2| Events that pass this selection criteria are treated as a separate signal region. Events with two or more jets that fail the VBF criteria but pass the more general described selection cuts in Section 7.5.1 are added to the ≥ 1 jet signal regions. 7.5.3 H → ττ results The total number of expected signal and background events in each H → ττ signal region is shown in Table 7.5.3. The signal yield in the ≥ 1 jet signal regions predominantly comes from the gluon-gluon fusion Higgs production mechanism. The MMC mass distribution for each of the seven H → ττ signal regions are shown in Figures 7.5.3 and 7.5.3. 0 jet Process Group e − τh μ − τh low pT miss high p T miss low p T miss high p T miss Z∕γ∗ → τ+τ− 3546 ±199 1344 ±125 7467 ±411 2772 ±256 W(→ ℓν) +jets 790 ±83 355 ±64 942 ±97 390 ±72 Z∕γ∗ → ℓℓ 1382 ±344 331 ±132 895 ±224 184 ±74 (Single) top 1.2 ±0.6 1.7 ±0.6 1.3 ±0.6 2.6 ±0.9 Diboson 7.4 ±1.1 4.6 ±1.8 9.5 ±1.1 6.3 ±1.4 QCD multi-jet 2738 ±105 516 ±36 1450.0 ±21 269 ±4 Total 8465 ±409 2553±194 10765 ±477 3623 ±275 Data 2011 8248 2512 10886 3563 Signal (mH = 120 GeV) 9.3 ±1.3 7.8 ±1.2 15.2 ±2.1 10.7 ±1.2 ≥ 1 jet VBF Process Group e − τh μ − τh ℓ − τh Z∕γ∗ → τ+τ− 1263 ±96 1843 ±133 45±4 W(→ ℓν) +jets 411 ±58 465 ±71 24±9 Z∕γ∗ → ℓℓ 355 ±92 90 ±32 12±5 (Single) top 168 ±12 172 ±12 12±0.9 Diboson 14.7 ±3.8 15.8 ±3.1 1.2±0.3 QCD multi-jet 354 ±20 120 ±5 20±2 Total 2567 ±145 2707 ±153 115±11 Data 2011 2574 2707 122 Signal (mH = 120 GeV) 9.8 ±2.0 12.3 ±2.5 3.0±0.4 (a) e − τ, 0 jet, low pmissT (b) μ − τ, 0 jet, low pmissT (c) e − τ, 0 jet, high pmissT (d) μ − τ, 0 jet, high pmissT (a) e − τ, ≥ 1 jet (b) μ − τ, ≥ 1 jet (c) ℓ − τ VBF region 7.6 Optimisation of the H →WW∗ →ℓντν event selection criteria Due to the different topology of H → WW∗ → ℓντν events relative to H → ττ events, it is necessary to apply separate selection criteria. In this section, the optimisation method described in Section 7.5 is employed using variables designed to exploit the small angular separation expected between the τh and light lepton. As in Section 7.5.1, the W+jets and QCD background yields for each set of cuts is estimated using the data-driven methods described in Section 7.7 and Section 7.8, respectively. The variables used in this optimisation included: the pmiss T of the event (Figure 7.6), on which the cut was varied from pmiss T > 0 GeV to pmiss T > 40 GeV in 10 GeV steps; the transverse mass of the event (defined in Equation 7.11) was varied from mT > 30 GeV to mT > 50 GeV in 10 GeV steps; and the angular separation of the light lepton and τh, defined as (7.13) and shown in Figure 7.6. The cut was varied from ΔR(ℓ,τh) < 2.05 to ΔR(ℓ,τh) < 1 in steps of 0.15. (a) e − τ (b) μ − τ (a) e − τ events (b) μ − τ events 7.6.1 H → WW∗ → ℓντν event selection cuts As a result of the H → WW∗ → ℓντν signal optimisation with m H = 125 GeV, events were required to pass the following selection criteria: Exactly one tightly identified τh and exactly one identified, isolated light lepton candidate. The candidates must be of opposite charge qτh × qℓ < 0. mT > 30 GeV. pmiss T > 10 GeV. ΔR(ℓ,τh) < 1.15. Due to the different lepton pT cuts, events with a final state e−τh pair and μ−τh pair are treated as separate signal regions. The total numbers of expected signal and background events in each H → WW∗ → ℓντν signal region are shown in Table 7.2 and the visible mass distribution for both of the H → WW∗ → ℓντν signal regions are shown in Figure 7.6.1. Process Group e − τ μ − τ W(→ ℓν) +jets 598±71 757±77 (single) top 69 ±4 69± 3 Z → ℓℓ 35±12 57 ±18 Z → ττ 32±10 33±8 Diboson 7.2±1.4 8.6±1.1 QCD multi-jet 19.1±4.7 12.2 ±0.9 Total 760±73 937±80 Signal (mH = 125 GeV) 1.1±0.1 1.4±0.2 Signal (mH = 160 GeV) 13.1±0.8 13.9±0.9 Data 2011 713 852 (a) e − τ (b) μ − τ 7.7 Data-driven estimate of the W(→ℓν) +jets background 7.7.1 Introduction To correctly model the yield of background events in which a hadronic jet has been mis-identified as a τh, i.e. the W(→ ℓν) +jets background, we cannot rely on Monte Carlo alone since the hadronisation of partons is not generally well modelled. Therefore, we construct regions of phase space that should be relatively free of signal and dominated by events arising from the W(→ ℓν) +jets process. By comparing the expected yield of W(→ ℓν) +jets from simulation in this region with the observed yield from data, a correction factor or scale factor (fW ) can be calculated and used to normalise the yield of these events in the signal regions. For each signal region, (defined in Sections 7.5.1 and 7.6.1) the different kinematic selection requirements on, for example, the electron and muon transverse momenta, will cause differences in the expected ratios of jets originating from quarks or gluons in W(→ ℓν) +jets events. Since jets originating from quarks are more likely to be mis-identified as a τh than jets originating from gluons, it is necessary to derive a separate fW scale factor for each signal region. It is also necessary to derive separate fW for events with the reconstructed charge product qℓ × qτh < 0 [opposite sign (OS) events] and qℓ × qτh > 0 [same sign (SS) events], since the fraction of mis-identified τh initiated by a quark or gluon is also different in these two regions. 7.7.2 W(→ ℓν) +jets in the H → ττ analysis The transverse mass is defined in Equation 7.11 and shown for four signal regions in Figure 7.5. Since the H → ττ signal peaks at mT ∼ 0 GeV, a W(→ ℓν) +jets control region is constructed around the W mass end-point (mT ∼ mW ). For each signal region, fW is defined as (7.14) where Ndata(70 < mT < 110 GeV), Nnon−W+jetsMC(70 < m T < 110 GeV) and NW+jetsMC(70 < m T < 110 GeV) are the yield of data, non W+jets events and W+jets events passing all the H → ττ selection criteria in a given region passing except for the transverse mass which is required to lie between 70 GeV and 110 GeV and second and third quantities are derived from MC. These scale factors are shown in Table 7.3. The differences in fW are due to the different fraction of jets originating from a quark or gluon in the sample due to the different light-lepton pT cut; the different jet multiplicities; the OS and SS regions quark fraction difference. SS/OS region Final state particles Signal Region fW OS e − τ low pT miss, 0 jet 0.533 ± 0.070 high pT miss, 0 jet 0.582 ± 0.006 ≥ 1 jet 0.610 ± 0.011 μ − τ low pT miss, 0 jet 0.431 ± 0.064 high pT miss, 0 jet 0.541 ± 0.005 ≥ 1 jet 0.604 ± 0.010 ℓ − τ VBF cuts 1.000 ± 0.123 SS e − τ low pT miss, 0 jet 1.562 ± 0.483 high pT miss, 0 jet 0.850 ± 0.017 ≥ 1 jet 0.781 ± 0.024 μ − τ low pT miss, 0 jet 0.793 ± 0.256 high pT miss, 0 jet 0.697 ± 0.012 ≥ 1 jet 0.726 ± 0.021 ℓ − τ VBF cuts 1.336 ± 0.263 Due to the large statistical uncertainties on the scale factors for the low pmiss T , 0-jet region, the high pmiss T , 0-jet region scale factors are used to normalise the yield of W+jets in all 0-jet regions. 7.7.3 W(→ ℓν) +jets in the H → WW∗ → ℓντν analysis For the H → WW∗ → ℓντν analysis, there is no such separation of signal and W(→ ℓν) +jets events in the transverse mass distributions. This motivates a different definition of the W(→ ℓν) +jets control region for the H → WW∗ → ℓντν analysis. (a) e − τ events (b) μ − τ events The signal and background Δϕ(ℓ,τh) distribution of events passing the object selection cuts and with mT > 30 GeV are shown in Figure 7.7.3. In order to construct a control region that is suitably pure, events are first required to have Δϕ(ℓ,τh) < 2. This selects a region dominated by W(→ ℓν) +jets production though this is also where the signal peaks. To isolate the W(→ ℓν) +jets from the signal, a cut is placed on the event’s ΔR(ℓ,τh) (Figure 7.6). The region ΔR(ℓ,τh) > 1.15 and Δϕ(ℓ,τh) < 2 is used as a control region to obtain a W(→ ℓν) +jets scale factor: (7.15) where Ndata(ΔR(ℓ,τh) > 1.15), Nnon−W+jetsMC(ΔR(ℓ,τ h) > 1.15) and NW+jetsMC(ΔR(ℓ,τ h) > 1.15) are the yield of data, non W(→ ℓν) +jets events and W(→ ℓν) +jets events after all other H → WW∗ → ℓντν selection criteria. The f W scale factors for e−τh and μ−τh events are shown in Table 7.7.3. Region Channel fW OS e − τh 0.618 ± 0.014 μ − τh 0.555 ± 0.008 SS e − τh 1.01 ± 0.10 μ − τh 0.71 ± 0.08 7.7.4 Systematic uncertainty on the W(→ ℓν) +jets data-driven background method A systematic uncertainty on the yield of W(→ ℓν) +jets in each signal region of the H → ττ and H → WW∗ → ℓντν analysis due to this method is defined by varying the f W for each region by its statistical uncertainty. These are shown in Tables 7.3 and 7.7.3 for the H → ττ and H → WW∗ → ℓντν analyses, respectively. 7.8 Data-driven estimate of the QCD multi-jet background 7.8.1 Introduction Simulated event samples cannot be used to model multi-jet backgrounds from QCD processes with several outgoing partons due to the very large multi-jet process cross sections. Instead, the data are used to estimate the contribution of QCD events using a so-called ‘ABCD’ method. In this method, the charge product of the light lepton and the τh, qτh × qℓ, and the isolation requirements on the light lepton are used to split events into four regions as shown in Table 7.5. The signal region (A) has an isolated, light lepton and qτh × qℓ < 0. The control regions (B, C and D) are expected to contain a negligible number of signal events and hence are modelled as containing events originating only from QCD and other background processes. Light lepton isolation Isolated Not isolated qℓ × qτh < 0 (OS) A (Signal) C qℓ × qτh > 0 (SS) B D To estimate the yield of QCD events in each signal region, first the expected number of events from other background processes (Z+jets, W+jets, diboson and tt and single top) is calculated for regions B, C and D. For each region, this expected yield is subtracted from the observed number of data events to obtain the number of QCD events in each region, nC and nD respectively. A ratio is then taken to find the OS/SS ratio of non-isolated QCD events, nC∕nD. To model the yield distribution shapes of QCD events in region A, region B is used, after normalising to the OS/SS ratio nC∕nD. 7.8.2 QCD background in the H → ττ analysis The nC∕nD ratio defined in Section 7.8 for each H → ττ signal region is shown in Table 7.8.2. Figure 7.8.2 shows that these ratios have no significant dependence on the calorimeter isolation requirement for all events passing the tau selection requirements. As with the fW determination in Section 7.7.2, the variations present are due to the different quark fraction of jets in each region. Final State Signal Region Non-isolated OS/SS ratio e − τ 0 jet, low pT miss 1.070 ± 0.022 0 jet, high pT miss 1.069 ± 0.039 ≥ 1 jet 1.025 ± 0.028 μ − τ 0 jet, low pT miss 1.154 ± 0.009 0 jet, high pT miss 1.154 ± 0.018 ≥ 1 jet 1.141 ± 0.011 ℓ − τ VBF cuts 1.209 ± 0.055 7.8.3 QCD background in the H → WW∗ → ℓντν analysis The multi-jet background yield and shape are estimated in the same way for the H → WW∗ → ℓντν signal regions as for the H → ττ signal regions, with the n C∕nD ratios shown in Table 7.8.3. After the Δϕ(ℓ,τh) requirement, there are very few events expected from QCD processes, since in a di-jet event two highest pT jets are expected to be back-to-back in the transverse plane. Signal Region Non-isolated OS/SS ratio e − τ 1.80 ± 0.44 μ − τ 1.41 ± 0.10 7.8.4 Systematic uncertainty on the QCD multi-jet data-driven background method A systematic uncertainty on the yield of QCD multi-jet events in each signal region of the H → ττ and H → WW∗ → ℓντν analysis from this method is defined by varying the non-isolated SS/OS ratio (nC∕nD) in each region by its statistical uncertainty. These uncertainties for each signal region of the H → ττ analysis are summarised in Table 7.8.2, while those calculated for the H → WW∗ → ℓντν analysis signal regions are shown in Table 7.8.3. An additional uncertainty could have been defined by using MC samples with a different parton showering model to determine the yield of non-QCD events in the control regions nB,nC and nD and taking the difference in the expected yield as a systematic uncertainty. Since the non-QCD yield of events in these regions is very small, the effect on the expected QCD multi-jet yield in the signal region is assumed to be negligible. 7.9 Systematic uncertainties Due to the imperfect modelling of both the signal and background physics processes and the simulation of the ATLAS detector response, small data-based corrections are applied and systematic uncertainties are determined. 7.9.1 Theory uncertainties In both analyses, the expected number of signal events are derived from simulated samples and scaled according to their next-to-next-to-leading-order (NNLO) cross sections listed in Appendix A.1 and the integrated luminosity of the collected data. All background samples (except for W(→ ℓν) +jets and QCD multi-jet production) are scaled by the NLO cross sections listed in Appendix A.1 and the integrated luminosity. In the NNLO Higgs production and NLO background calculations, the cross section for each simulated process (Appendix A.1) was re-calculated with the factorisation and renormalisation scales halved and doubled simultaneously, with the maximum variation of the resulting cross section in either direction taken as a systematic uncertainty. The parton density function systematic uncertainty was assigned by taking the difference in cross section when using a different PDF set [75]. Since the change in the visible mass and MMC mass distribution shapes from these variations is negligibly small, only the overall normalisation is treated as a source of systematic uncertainty. Since the W(→ ℓν) +jets yield is normalised to data there is no need to account for an uncertainty on the overall cross-section. The dominant theory uncertainties on the main backgrounds are summarised in Table 7.9.1. The theory uncertainties on the signal processes as a function of mH are shown in Appendix A.1; the gluon-gluon fusion process uncertainties sum in quadrature to be ≃ 18% while the vector-boson-fusion process uncertainties sum in quadrature to be ≃ 4% [14]. Process QCD scale (%) PDF variation (%) q2 scale (%) Z → ℓℓ∕ττ 1.0 2.0 12.5 tt 1.0 8.0 3.0 The systematic uncertainty on the integrated luminosity of the data samples is estimated to be 3.9% by comparing the integrated luminosity calculated from various techniques, as described in reference [72]. 7.9.2 Trigger efficiencies The measured trigger efficiency scale factors is defined to be the observed difference between the MC expectation of the trigger efficency and data. These were calculated using Z → ℓℓ data with tag-and-probe methods. The electron and muon triggered events are shown in Table 7.9.2. Trigger Scale Factor (%) Muon (pT > 18 GeV) 99.2 ± 0.5 Electron (pT > 20(22) GeV) 99.5 ± 1.0 7.9.3 Electron candidates For electrons, systematic uncertainties on the following are considered: energy scale and resolution, reconstruction and identification efficiency, and calorimeter isolation efficiency [74]. Energy clusters in the electromagnetic calorimeter identified as resulting from primary electrons have an energy scale uncertainty of about 1% (3%) in the barrel (endcap) regions, as determined by studies that calibrate data samples using resonances that decay to electrons, such as Z → e+e− and an estimate of energy deposited in the upstream material. The electron identification efficiency and its associated systematic uncertainty are evaluated using tag-and-probe methods using electrons produced in the decay of J∕ψ and Z resonances, and the uncertainty is found to be about 1.5%, varying with the candidate ET and η. The uncertainty on the e − τh event selection is obtained by varying the identification efficiency by this amount. The uncertainty on the electron calorimeter isolation scale factor (2%) is found using Z → e+e− events in which a comparison of the electron isolation efficiency for data and simulation is made. The effect of the incertainty is determined by varying the isolation scale factor by its uncertainty. 7.9.4 Muon candidates Systematic uncertainties on the muon pT resolution and the identification efficiency scale factor are applied by varying these quantities by their uncertainties [73]. The muon pT resolution uncertainty is obtained by calibrating the simulated events to data using the measured Z boson peak width, while the identification efficiency is measured using a tag-and-probe analysis of muons produced in Z → μ+μ− events in data. 7.9.5 τh candidates The τh energy scale systematic uncertainty [58] is based on studies of simulated Z → ττ events. In these events the relative difference of the reconstructed τh pT and that of the hadronic tau decay in the Monte Carlo truth record is studied using reconstructed, identified tau candidates within a cone of radius ΔR = 0.2 from a hadronic tau decay in the MC truth record. The sources of uncertainty considered include variations in the hadronic shower model used, energy cluster noise thresholds, and the amount of additional dead material traversed before reaching the calorimeters in the simulated data. A systematic uncertainty is also assigned to a scale factor for the mis-identification rate of an electron as a tau candidate [58], measured using Z → e+e− data events with a reconstructed me+e− within a narrow window around mZ. This scale factor is only used for simulated events in which the reconstructed τh candidate is matched to a simulated primary electron with pT > 5 GeV within a cone of radius ΔR = 0.2. 7.9.6 Jet candidates Systematic uncertainties arising from the estimation of the jet energy scale in [77, 78] are taken into account. The jet energy scale is measured at the electromagnetic (EM) scale and is calibrated using energy deposits in the calorimeters from electromagnetic showers. This energy is established using test-beam measurements for electrons in the barrel and endcap calorimeters. Corrections to this are applied to account for the hadronic calorimeter response as well as energy loss in dead material derived from Monte Carlo simulated events that restore the calorimeter response of the reconstructed jet to the simulated jet response. 7.9.7 pmissT systematic uncertainties Since both analyses have a cut on the pmiss T of the event, the energy scale variation of the jets, electrons and τh are propagated to the pmiss T vector and the effect on the acceptance of this cut is evaluated [79]. The energy scale variations of the clusters associated with a selected jet or τh candidate are treated as fully correlated. The effect of energy scale uncertainty on the acceptance is also evaluated for energy clusters not associated with a reconstructed object. 7.9.8 Data driven background estimation systematic uncertainties The systematic uncertainties associated with these methods are described in Sections 7.7.4 and 7.8.4. 7.9.9 Systematic uncertainties in the H → ττ analysis The systematic uncertainties arising from the object selection are shown in Tables 7.10 and 7.11 for the H → ττ signal regions for events with no additional selected jets and ≥ 1 additional selected jet in the final state, respectively. The W(→ ℓν) +jets normalisation is taken from the control region, but uncertainties on the W(→ ℓν) +jets yield account for any differences in the object selection efficiencies, energy scales or resolutions between the W control regions and the signal regions since the scale factors fW are normalised to data. 0 jet (pmiss T < 20 GeV) Energy Scale Electrons Muons τh pmiss T Jet/τh Electron Res. Id. Res. Id. Id. e-FR Cluster Pile-up ggF +3.0−4.3 +0.0+0.0 +0.1+0.2 −1.4+1.1 −0.1+0.1 −1.4+1.2 −4.1+4.1 −0.0+0.0 +8.4−9.6 +4.7−6.0                       VBF +15.0−24.4 −0.0+0.1 −0.0+1.0 −1.4+1.2 −0.0+0.0 −1.3+1.2 −4.2+4.2 −0.0+0.0 +12.3−19.5 +1.0−9.8                       Z → ττ −2.9+3.2 −0.4+0.3 −0.1−0.1 −1.2+1.0 −0.1+0.0 −1.5+1.3 −4.3+4.3 −0.0+0.0 −0.0−0.0 −0.0−0.0                       W(→ ℓν) +jets + jets −5.1+1.3 −0.3+0.0 −1.1+0.0 −0.0+0.0 −0.2+0.1 −0.1+0.1 +0.7−0.7 +0.5−0.5 −3.2+3.4 −2.9+1.4                       Z → ℓℓ −2.4+1.5 +0.4−0.2 +0.3+0.1 −2.0+1.6 −0.5+0.4 −1.0+0.9 −4.4+4.4 −22.6+22.6 +8.5−8.6 +5.2−5.0                       (single) top +10.3−20.3 −0.8+0.0 −1.3+0.0 −1.6+1.3 −5.7+4.7 −1.2+1.1 −4.6+4.6 −0.0+0.0 −0.0+12.7 −0.0+2.7                       Diboson +0.6−2.7 +0.2+0.2 −0.0+0.8 −1.5+1.3 −0.2+0.1 −1.3+1.1 −4.4+4.4 −1.6+1.6 −0.9+0.0 −0.4+0.0                       0 jet (high pmiss T ) Energy Scale Electrons Muons τh pmiss T Jet/τh Electron Res. Id. Res. Id. Id. e-FR Cluster Pile-up ggF −0.0+1.6 −0.4+0.3 −0.2−0.3 −1.5+1.2 −0.1+0.1 −1.3+1.2 −4.0+4.0 −0.0+0.0 −5.0+6.2 −2.0+4.5                       VBF +18.0−12.4 −0.8+0.0 +0.1−0.9 −1.4+1.1 −0.0+0.0 −1.4+1.2 −4.1+4.1 −0.0+0.0 −1.9+6.6 −0.0+5.1                       Z → ττ −8.2+8.4 −0.2+0.0 −0.1−0.1 −1.2+1.0 −0.1+0.1 −1.5+1.3 −4.1+4.1 −0.0+0.0 −0.0−0.0 +0.0−0.0                       W(→ ℓν) +jets + jets −6.3+0.0 −0.7+0.0 −1.8+0.0 −0.1+0.1 −1.5+1.2 −0.2+0.1 +1.0−1.0 +0.5−0.5 −17.8+6.2 −11.7+4.4                       Z → ℓℓ −5.8+1.0 −1.0−0.0 −0.3+0.0 −2.2+1.8 −0.9+0.6 −0.9+0.8 −4.3+4.3 −29.0+29.0 −16.5+19.9 −8.7+11.5                       (single) top +21.6−21.7 −2.6+0.0 −3.1+0.0 −1.5+1.2 −4.1+3.2 −1.3+1.2 −4.1+4.1 −0.0+0.0 −0.0+2.8 −0.0+5.1                       Diboson −2.9+0.0 +0.8−0.7 +0.4+0.3 −1.5+1.2 −0.2+0.2 −1.2+1.1 −4.2+4.2 −0.8+0.8 −6.9+6.5 −5.3+0.0                       ≥ 1 jet Energy Scale Electrons Muons τh pmiss T Jet/τh Electron Res. Id. Res. Id. Id. e-FR Cluster Pile-up ggF −10.0+9.8 +0.3−0.3 +0.1+0.1 −1.5+1.2 −0.0+0.0 −1.3+1.1 −4.2+4.2 −0.0+0.0 +1.3−0.2 +1.4−0.1                       VBF −2.3+0.9 +0.2−0.1 +0.0+0.0 −1.6+1.3 −0.1+0.1 −1.2+1.1 −4.3+4.3 −0.0+0.0 +0.5−1.9 +0.5−0.9                       Z → ττ −4.5+5.0 −0.1+0.2 +0.0+0.1 −1.5+1.2 −0.0+0.0 −1.3+1.2 −4.4+4.4 −0.0+0.0 −0.0+0.0 −0.0−0.0                       W → ℓν + jets −4.2+0.0 +1.1−2.1 −0.3+0.1 −0.5+0.4 −1.6+1.3 −0.5+0.4 +1.7−1.7 +0.7−0.7 −13.4+0.0 −13.0+0.0                       Z → ℓℓ −14.1+18.8 −2.0+0.0 +0.1−1.0 −2.8+2.3 −0.5+0.4 −0.5+0.4 −4.7+4.7 −14.1+14.1 −7.7+10.0 −5.4+5.8                       (single) top −3.9+4.2 −0.0+0.0 −0.4+0.0 −1.8+1.4 −0.1+0.0 −1.1+1.0 −4.4+4.4 −0.4+0.4 −0.7+0.6 −0.6+0.3                       Diboson −4.8+11.0 +0.9−0.1 +0.3+0.1 −1.7+1.4 −0.1+0.1 −1.2+1.0 −4.4+4.4 −0.7+0.7 −1.6+0.4 −0.4+0.0                       VBF signal region Energy Scale Electrons Muons τh pmiss T Jet/τh Electron Res. Id. Res. Id. Id. e-FR Cluster Pile-up ggF −19.7+26.0 −1.6+0.0 −0.0+0.0 −2.4+1.9 −0.0+0.0 −0.7+0.6 −4.4+4.4 −0.0+0.0 −0.0+6.2 −0.0+6.2                       VBF −9.2+9.9 −0.2+0.0 +0.1−0.3 −1.6+1.3 −0.5+0.4 −1.2+1.1 −4.2+4.2 −0.0+0.0 +0.0−1.9 −0.1+0.0                       Z → ττ −4.4+5.2 −0.0+0.0 −0.0+0.0 −1.3+1.1 −1.6+1.3 −1.4+1.2 −4.3+4.3 −0.0+0.0 −0.0+0.0 −0.0+0.0                       W → ℓν + jets +24.7−39.4 +0.1+0.1 +0.3+0.0 −0.9+0.8 −5.6+4.8 −0.5+0.4 +2.6−2.6 +1.4−1.4 −6.0+0.0 −3.2+4.8                       Z → ℓℓ −7.6+47.7 −3.9+0.0 −3.9+0.0 −3.2+2.6 −2.1+1.7 −0.2+0.2 −5.0+5.0 −2.0+2.0 −0.9+0.3 −2.6+0.0                       (single) top −5.2+2.2 −1.8+0.0 −2.2+0.0 −2.0+1.7 −2.0+1.8 −0.9+0.8 −4.5+4.5 −0.0+0.0 +0.4−2.0 −0.9+0.0                       Diboson −14.2+14.5 −0.0+0.2 −0.5+0.7 −1.3+1.1 −0.2+0.2 −1.4+1.2 −4.3+4.3 −0.6+0.6 −0.0+10.6 +6.7−3.7                       7.9.10 Systematic uncertainties in the H → WW∗ → ℓντν analysis The systematic uncertainties arising from the object selection are shown in Table 7.9.10 for the H → WW∗ → ℓντν signal regions. Process Energy Scale Electrons Muons τh pmiss T Jet∕τh Electron Res. Id. Res. Id. Id. e-FR Cluster Pile-up ggF −5.5+0.5 −0.6+0.9 0.0+0.4 −0.8+0.8 −0.1+0.0 −0.2+0.2 −4.2+4.2 −1.7+1.7 −0.4+0.9 −0.5+0.8                       VBF −5.5+0.5 −0.6+0.9 0.0+0.4 −0.8+0.8 −0.1+0.1 −0.2+0.2 −4.2+4.2 −1.7+1.7 −0.4+0.9 −0.5+0.8                       Z → ττ −0.6+0.0 −1.2+1.9 −0.2+0.9 −0.8+0.8 −0.4+0.3 −0.3+0.3 −4.2+4.2 −0.1+0.1 −4.1+5.4 −2.5+3.7                       W → ℓν + jets −6.2+0.0 −0.5+0.5 0.0+0.3 −0.8+0.8 −0.0+0.0 −0.2+0.2 −4.2+4.2 −0.1+0.1 −0.9+0.5 −0.5+0.4                       Z → ℓℓ + jets −4.3+0.0 −0.4+1.3 0.0+1.2 −0.8+0.8 −0.4+13.2 −0.3+0.3 −4.2+4.2 −16.8+16.8 −1.8+3.0 0.0+1.4                       (Single) Top −3.3+0.0 −0.2+0.5 0.0+0.2 −0.9+0.9 0.0+0.1 −0.3+0.3 −4.2+4.2 −0.6+0.6 −0.3+0.2 −0.2+0.0                       Diboson −1.6+0.0 −0.3+0.3 −0.0+0.4 −0.8+0.8 0.0+0.1 −0.2+0.2 −4.2+4.2 −1.3+1.3 −0.6+0.6 −0.0+0.5                       7.10 Limit Setting Since the number of observed data events in the H → ττ and H → WW∗ → ℓντν analysis channels shown in Tables 7.5.3 and 7.2 are consistent with a background only hypothesis, no evidence of Higgs boson production is observed (though it is not expected that either analysis is sensitive to SM Higgs boson production with this amount of integrated luminosity). An upper limit is placed on the excluded cross section of SM like Higgs boson production, as a function of the Higgs mass using the H → ττ and H → WW∗ → ℓντν signal regions separately. The limits are defined to have an exclusion at 95% confidence level and are calculated using the profile likelihood method [80]. The MMC mass and visible mass are used as the discriminating variables in the H → ττ and H → WW∗ → ℓντν signal regions, respectively. Systematic uncertainties are included as nuisance parameters. Systematic uncertainties on the shape and normalisation of the MMC and visible mass distributions due to the variation of the jet and τh energy scales are included. Other uncertainties are described in Section 7.9 including the measurement of integrated luminosity, object energy scales, resolutions and the acceptance common to all samples and signal regions are treated as fully correlated and constrained using Gaussian functions. The likelihood function for a particular bin of the MMC distribution in a given signal region is defined as: (7.16) where: P(a, b) is a Poisson distribution (7.17) n is the number of data events in a given bin. stat are the statistical uncertainties on the MC or data driven estimates. s,b are nuisance parameters (vectors of systematic uncertainties) for each signal and background process. global are the common systematic uncertainties that are fully correlated across all signal and background processes (e.g. the luminosity uncertainty). ( s, b, global) is a parameterisation of the nuisance parameters which are constrained using Gaussian functions (7.18) μT is the number of expected events, as defined by (7.19) In the definition of μT , L is the integrated luminosity, μ is the scaling factor of the expected SM signal cross section (signal strength) and μ = 0 (1) corresponds to the absence (presence) of a SM Higgs boson signal, σl(mH) is the effective cross section for signal events in each SM Higgs production process, l = ggF,VBF,WH and ZH, βj is the cross section for background process j, and fs,b( s,b) is the dependence of the expected number of events on each nuisance parameter. A test statistic is obtained from the profile-likelihood ratio using asymptotic formulae [80], as given by (7.20) where: is the maximum likelihood estimator of μ, represents the nuisance parameters evaluated at μ, (μ) represents the maximum likelihood estimators of at a given μ. Pseudo experiments are generated using the distributions of signal and background to obtain a PDF f( μ,μ,(μ)) for a given signal strength, μ. Using this PDF, a p-value (the probability that a background-only experiment fluctuates more than the observation) for μ is obtained using: (7.21) An upper limit is set on the signal strength μ by iteratively evaluating this integral until pμ = 0.05. Similarly, pseudo experiments using background only distributions are used to evaluate an expected upper limit on μ, along with ±1 and ±2 standard deviation (SD) values at each assumed signal mass. 7.10.1 Limit The expected and observed upper limits on μ as a function of assumed SM Higgs boson mass is shown in Figure 7.10.1 for the H → ττ analysis. As mH increases, the SM Higgs production cross section and (H → ττ) both fall, leading to a higher expected upper limit at higher mH. Conversely, the primary background process (Z → ττ) becomes less important as mH increases due to the ττ mass resolution obtained using the MMC method. The shape of the expected upper limit as a function of Higgs mass is determined by these two effects. The expected and observed upper limits on μ as a function of assumed SM Higgs boson mass is shown in Figure 7.10.1 for the H → WW∗ → ℓντν analysis. As m H increases, (H → WW∗) increases faster than the SM Higgs production cross section decreases, leading to a lower expected upper limit. The dominant uncertainty is due to the uncertainty on the primary background, W+jets production with a fake τh, due to the limited W+jets MC sample size and the uncertainty on the W+jets scale factor, fW . Chapter 8 Conclusions and outlook In the Standard Model (SM) of particle physics, a mechanism is required to generate the observed masses of the SM fermions and electroweak gauge bosons. The discovery of Higgs boson production at the LHC would provide evidence for the so-called Higgs mechanism. A search for a light SM Higgs boson in the H → ττ and H → WW∗ → ℓντν decay modes using 4.7 fb−1 of data collected at the ATLAS detector in 2011 has been presented. The number of events passing an event selection is consistent with total background estimation from other SM processes and an observed upper limit is placed on the SM Higgs boson production rate at 95% confidence level at around 5 times the SM cross section in the range 100 < mH < 130 GeV∕c2. Global fits of indirect electroweak measurements combined with the results of previous direct searches favour a relatively light Higgs boson with a mass 114 < mH < 157 GeV∕c2. In this mass range, a search has been made using one of the most sensitive SM Higgs decay channels, the decay to tau leptons. This result was combined with other ATLAS SM Higgs searches [81] to exclude a SM Higgs boson with 110 < mH < 117.5 GeV∕c2, 118.5 < m H < 122.5 GeV∕c2 and 129 < mH < 539 GeV∕c2 at 95% confidence level while the range 120 < m H < 555 GeV∕c2 was expected to have been excluded. With a larger data sample, there are several possible ways in which to improve the sensitivity of the H → ττ and H → WW∗ → ℓντν analyses beyond the naive improvement due to increased statistics alone. By further specialisation of the signal regions, cuts can be designed to further improve the signal to background ratio. For example, it may be possible to design cuts to select the associated production modes W∕Z(→ jet jet)H(→ ττ) or events in which the Higgs boson is boosted in recoil against a jet. Another way would be to move to a multi-variate based discriminant such as a Boosted Decision Tree or a Neural Network in both analyses to exploit differences in the correlations of the signal and background event kinematics. In 2012, it is expected that ∼ 20 fb−1 will be collected at = 8 TeV by the ATLAS detector. With this increased sample size, it is expected that the combined results of the Higgs analyses proceeding at the ATLAS experiment should be able to either exclude at 95% confidence level or claim discovery of a light, SM Higgs boson. Appendix A H →ττ and H →WW∗→ℓντν analysis appendix A.1 Simulated Datasets The cross section at = 7 TeV used to normalise the expected yield of each process considered as background to the H → ττ and H → WW∗ → ℓντν analyses described in Chapter 7 are listed in Table A.1, Table A.2 and Table A.3 along with the generator used for that process. The NNLO Higgs production cross section at = 7 TeV for the gluon gluon fusion (ggF) and weak boson fusion (VBF) simulated samples along with the branching ratios (BR) used for H → ττ and H → WW∗ are shown as a function of Higgs mass in Table A.1. The systematic uncertainties on the cross section under variation of αs, PDF set and factorisation and renormalisation scales is shown in Table A.1. Process σ ×× k − factor × ϵfilter (pb) Nevents Generator Z → τ+τ−→ ℓτh+Np0 835.5 10613180 Alpgen Z → τ+τ−→ ℓτh+Np1 167.95 3334138 Alpgen Z → τ+τ−→ ℓτh+Np2 50.45 804948 Alpgen Z → τ+τ−→ ℓτh+Np3 14.06 509848 Alpgen Z → τ+τ−→ ℓτh+Np4 3.49 145000 Alpgen Z → τ+τ−→ ℓτh+Np5 0.96 45001 Alpgen Z∕γ∗→ τ+τ−→ ℓτh+Np0 (10 < mll < 40) 3727.22 875000 Alpgen Z∕γ∗→ τ+τ−→ ℓτh+Np1 (10 < mll < 40) 103.61 300000 Alpgen Z∕γ∗→ τ+τ−→ ℓτh+Np2 (10 < mll < 40) 50.51 399000 Alpgen Z∕γ∗→ τ+τ−→ ℓτh+Np3 (10 < mll < 40) 10.2 150001 Alpgen Z∕γ∗→ τ+τ−→ ℓτh+Np4 (10 < mll < 40) 2.26 40000 Alpgen Z∕γ∗→ τ+τ−→ ℓτh+Np5 (10 < mll < 40) 0.56 10001 Alpgen Z → e+e−+Np0 835.4 6218285 Alpgen Z → e+e−+Np1 167.95 1234998 Alpgen Z → e+e−+Np2 50.68 810000 Alpgen Z → e+e−+Np3 13.95 220001 Alpgen Z → e+e−+Np4 3.6 60001 Alpgen Z → e+e−+Np5 1.04 50001 Alpgen Z → μ+μ−+Np0 835.4 6615231 Alpgen Z → μ+μ−+Np1 167.68 1334297 Alpgen Z → μ+μ−+Np2 50.41 304948 Alpgen Z → μ+μ−+Np3 13.99 110001 Alpgen Z → μ+μ−+Np4 3.44 30001 Alpgen Z → μ+μ−+Np5 0.96 10001 Alpgen Z∕γ∗→ e+e−+Np0 (10 < mll < 40) 3727.1 994950 Alpgen Z∕γ∗→ e+e−+Np1 (10 < mll < 40) 103.6 299999 Alpgen Z∕γ∗→ e+e−+Np2 (10 < mll < 40) 50.51 799950 Alpgen Z∕γ∗→ e+e−+Np3 (10 < mll < 40) 10.22 149999 Alpgen Z∕γ∗→ e+e−+Np4 (10 < mll < 40) 2.26 40001 Alpgen Z∕γ∗→ e+e−+Np5 (10 < mll < 40) 0.56 10001 Alpgen Z∕γ∗→ μ+μ−+Np0 (10 < mll < 40) 3727.1 999850 Alpgen Z∕γ∗→ μ+μ−+Np1 (10 < mll < 40) 103.54 300001 Alpgen Z∕γ∗→ μ+μ−+Np2 (10 < mll < 40) 50.57 499998 Alpgen Z∕γ∗→ μ+μ−+Np3 (10 < mll < 40) 10.22 150001 Alpgen Z∕γ∗→ μ+μ−+Np4 (10 < mll < 40) 2.26 40000 Alpgen Z∕γ∗→ μ+μ−+Np5 (10 < mll < 40) 0.56 10001 Alpgen Process σ ×× k − factor × ϵfilter (pb) Nevents Generator W → eν +Np0 8245.2 3358886 Alpgen W → eν +Np1 1551.6 2199646 Alpgen W → eν +Np2 451.92 3768633 Alpgen W → eν +Np3 121.56 908948 Alpgen W → eν +Np4 30.3 250001 Alpgen W → eν +Np5 8.55 70000 Alpgen W → μν +Np0 8245.2 3462943 Alpgen W → μν +Np1 1551.6 2348645 Alpgen W → μν +Np2 451.92 3768738 Alpgen W → μν +Np3 121.56 1008447 Alpgen W → μν +Np4 30.3 254951 Alpgen W → μν +Np5 8.55 70001 Alpgen W → τν +Np0 8245.2 3358886 Alpgen W → τν +Np1 1551.6 2249196 Alpgen W → τν +Np2 451.92 3750987 Alpgen W → τν +Np3 121.56 1009947 Alpgen W → τν +Np4 30.3 249999 Alpgen W → τν +Np5 8.55 65001 Alpgen Process σ ×× k − factor × ϵfilter (pb) Nevents Generator W−Z → lνll 0.09 100001 MC@NLO W−Z → lνqq 0.92 25001 MC@NLO W−Z → lνττ 0.04 25001 MC@NLO W−Z → qq′ll 0.27 100001 MC@NLO W−Z → qq′ττ 0.26 25001 MC@NLO W−Z → τνll 0.04 25001 MC@NLO W−Z → τνττ 0.02 25001 MC@NLO W+Z → lνll 0.16 25001 MC@NLO W+Z → lνqq 1.7 25001 MC@NLO W+Z → lνττ 0.08 25001 MC@NLO W+Z → qq′ll 0.51 24951 MC@NLO W+Z → qq′ττ 0.26 25001 MC@NLO W+Z → τνll 0.04 25001 MC@NLO W+Z → τνττ 0.02 24951 MC@NLO ZZ → 2l2τ 0.03 25001 MC@NLO ZZ → 4τ 0.01 25001 MC@NLO ZZ → llll 0.03 50001 MC@NLO ZZ → llνν 0.15 100000 MC@NLO ZZ → llqq 0.53 25001 MC@NLO ZZ → ττνν 0.08 25001 MC@NLO ZZ → ττqq 0.27 25001 MC@NLO qq′→ W+W−→ eνeν 0.52 199950 MC@NLO qq′→ W+W−→ eνμν 0.52 200001 MC@NLO qq′→ W+W−→ eντν 0.52 200001 MC@NLO qq′→ W+W−→ μνeν 0.52 199950 MC@NLO qq′→ W+W−→ μνμν 0.52 199001 MC@NLO qq′→ W+W−→ μντν 0.52 100001 MC@NLO qq′→ W+W−→ τνeν 0.52 199951 MC@NLO qq′→ W+W−→ τνμν 0.52 200001 MC@NLO qq′→ W+W−→ τντν 0.52 199677 MC@NLO gg → W+W−→ eνeν 0.02 10001 gg2WW gg → W+W−→ eνμν 0.02 10001 gg2WW gg → W+W−→ eντν 0.01 10001 gg2WW gg → W+W−→ μνeν 0.02 10001 gg2WW gg → W+W−→ μνμν 0.02 10000 gg2WW gg → W+W−→ μντν 0.01 10001 gg2WW gg → W+W−→ τνeν 0.01 10001 gg2WW gg → W+W−→ τνμν 0.01 10001 gg2WW gg → W+W−→ τντν 0.01 10001 gg2WW tt (No fully hadronic decays) 91.34 7146746 MC@NLO tt (fully hadronic decays) 73.23 1199035 MC@NLO single top : s − channel W → eν 0.5 99901 AcerMC single top : s − channel W → μν 0.5 199851 AcerMC single top : s − channel W → τν 0.5 175001 AcerMC single top : t − channel W → eν 7.83 999949 AcerMC single top : t − channel W → μν 7.83 999949 AcerMC single top : t − channel W → τν 7.83 998996 AcerMC single top : Wt − channel 15.6 769898 AcerMC mH (GeV) ggF σ(pp → H) (pb) VBF σ(pp → qq′H) (pb) BR(H → ττ) BR(H → WW∗) 100 24.02 1.546 0.0836 0.0111 105 21.78 1.472 0.0825 0.0243 110 19.84 1.398 0.0802 0.0482 115 18.13 1.332 0.0765 0.0867 120 16.63 1.269 0.0710 0.143 125 15.31 1.211 0.0637 0.216 130 14.12 1.6868 0.0548 0.305 135 13.08 1.10 0.0452 0.403 140 12.13 1.052 0.0354 0.504 145 11.27 1.004 0.0261 0.603 150 10.50 0.9617 0.0178 0.699 160 9.080 0.8787 0.00396 0.909 170 7.729 0.8173 0.000920 0.965 180 6.739 0.7480 0.000587 0.932 mH (GeV) q2 variation (%) PDF + α s variation (%) ggF VBF ggF VBF 100 16.5 1.0 8.0 4.0 105 16.4 1.0 8.0 4.0 110 16.3 1.0 8.0 4.0 115 16.5 1.0 8.9 4.0 120 16.3 1.0 8.0 4.0 125 16.3 1.0 8.0 4.0 130 16.4 1.0 8.0 4.0 135 16.1 1.0 8.0 4.0 140 16.2 1.0 8.0 4.0 145 16.4 1.0 8.0 4.0 150 16.6 1.0 8.0 4.0 160 16.6 1.0 8.0 4.0 170 16.6 1.0 8.0 4.0 180 16.6 1.0 8.0 4.0 References [1] S. L. Glashow. Partial Symmetries of Weak Interactions. Nucl. Phys., 22:579–588, 1961. [2] Steven Weinberg. A model of leptons. Phys. Rev. Lett., 19:1264–1266, Nov 1967. [3] A. Salam and J. C. Ward. Electromagnetic and weak interactions. Phys. Rev. Lett., pages 168–171, 1964. [4] Y. Fukuda et al. Evidence for oscillation of atmospheric neutrinos. Phys. Rev. Lett., 81:1562–1567, Aug 1998. [5] F. Halzen and A. D. Martin. Quarks and Leptons: An Introductory Course in Modern Particle Physics. Wiley, 1984. [6] K. et al Nakamura. Review of Particle Physics. Journal of Physics G, 33:1+, 2006. [7] C. Amsler et al. Review of particle physics. Physics Letters B, 667(15):1 – 6, 2008. Review of Particle Physics. [8] P.W. Higgs. Phys. Lett., 12:132, 1964. [9] F. Englert and R. Brout. Phys. Rev. Lett., 13:321, 1964. [10] P.W. Higgs. Phys. Rev. Lett., 13:508, 1964. [11] P.W. Higgs. Phys. Rev. Lett., 145:1156, 1966. [12] LEP Electroweak Working Group. . 2011. http://lepewwg.web.cern.ch/LEPEWWG/. [13] Tevatron New-Phenomena and Higgs Working Group. . hep-ex/1203.3774, 2011. [14] LHC Higgs Cross Section Working Group, S. Dittmaier, C. Mariotti, G. Passarino, and R. Tanaka (Eds.). Handbook of LHC Higgs Cross Sections: 1. Inclusive Observables. CERN-2011-002, CERN, Geneva, 2011. [15] The CMS Collaboration. G. l. bayatian et al. J. Phys. G: Nucl. Part. Phys., 34(995), 2007. [16] S. Asai et al. Eur. Phys. J. C, 32S2(19), 2004. [17] D. Rainwater F. Maltoni and S. Willenbrok. Phys. Rev. D, 66(034022), 2002. [18] E. Richter-Was and Acta. M. Sapinski. Phys. Polon. B, 30(1001), 1999. [19] T. Müller V. Drollinger and D. Denegri. arXiv:hep-ph/0111312v1 (2001). [20] M. Rubin J. M. Butterworth, A. R. Davison and G. P. Salam. Phys. Rev. Lett., 104(111801), 2010. [21] K. Black et al. arXiv:1010.3698v1, 2010. [22] E. Gabrielli et al. Nucl. Phys. B, 64(781), 2007. [23] A. Nikitenko D. Zeppenfeld, R. Kinnunen and E. Richter-Was. Phys. Rev. D, 62(013309), 2000. [24] M. Duhrssen et al. Phys. Rev. D, 70(113009), 2004. [25] L. Evans and P. Bryant. LHC machine. Journal of Instrumentation, 3:S08001, 2008. [26] C. Lefèvre (CERN). CERM-DI-0812015, 2008. [27] G. Aad et al. The ATLAS experiment at the CERN large hadron collider. Journal of Instrumentation, 3:S08003, 2008. [28] G. Aad et al. Expected performance of the ATLAS experiment: detector, trigger and physics. arXiv, 901:2008–20. [29] R. S. Thorne A. D. Martin, W. J. Stirling and G. Watt. . http://projects. hepforge.org/mstwpdf/. [30] ATLAS: Detector and physics performance technical design report. Volume 1. CERN-LHCC-99-14. [31] R. Brun and F. Rademakers. ROOT–An object oriented data analysis framework. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 389(1-2):81–86, 1997. [32] Matteo Cacciari, Gavin P. Salam, and Gregory Soyez. The anti-kt jet clustering algorithm. JHEP, 04:063, 2008. [33] The ATLAS Collaboration. Expected Performance of the ATLAS Experiment, Detector, Trigger and Physics, volume CERN-OPEN-2008-020. CERN, Geneva, 2009. Pages 368-396. [34] The ATLAS Collaboration. Commisioning of the atlas high-performance b-tagging algorithms in 7 tev data. 2011. ATLAS-CONF-2011-102. [35] The ATLAS Collaboration. Eur. Phys. J. C, 72:1909, 2012. [36] The ATLAS Collaboration. Eur. Phys. J. C, 72:1844, 2012. [37] S. Agostinelli et al. The GEANT4 Collaboration. Geant4 - a simulation toolkit. Nucl. Instrum. Meth, A506:250, 2003. [38] A. Barr and C.Boddy. Phys. Educ., 47:270–273, 2012. [39] Google Ic. 2011 Android. . http://www.android.com. [40] Gary Taylor. Visualizing the atlas inner detector with atlantis. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 549. [41] OpenGL-ES. 2011. . http://www.khronos.org/opengles. [42] A. Djouadi. Phys. Rep., 457(1), 2008. [43] S. Peñaranda J. Guasch, W. Hollik. Phys. Lett. B, 367(505), 2001. [44] T. Gleisberg et al. J. High Energy Phys., 007(0902), 2009. We use Version 1.2.1. [45] Pavel M. Nadolsky et al. Implications of CTEQ global analysis for collider observables. Phys. Rev., D78:013004, 2008. [46] F. Piccinini R. Pittau A. Polosa M.L. Mangano, M. Moretti. Alpgen, a generator for hard multiparton processes in hadronic collisions. JHEP, 07:001, 2003. [47] Torbjorn Sjostrand, Stephen Mrenna, and Peter Skands. A Brief Introduction to PYTHIA 8.1. Comput. Phys. Commun., 178:852–867, 2008. [48] M. Bähr et al. Eur. Phys. J. C, 58(639), 2008. [49] X. Rouby S. Ovyn and V. Lemaitre. arXiv:hep-ph/0903.2225v3, 2010. [50] M. Cacciari and G. P. Salam. Phys. Lett. B, 57(641), 2006. [51] The ATLAS Collaboration. Measurement of the mis-identification probability of tau leptons from hadronic jets and from electrons. 2011. ATLAS-CONF-2011-113. [52] M. Spira. http://people.web.psi.ch/spira/v2hv/. [53] J. Kalinowkski A. Djouadi and M. Spira. Comput. Phys. Commun., 56(108), 1998. [54] J. Campbell, R.K. Ellis, and D. Rainwater. Next-to-leading order QCD predictions for W+ 2 jet and Z+ 2 jet production at the CERN LHC. Physical Review D, 68(9):094021, 2003. [55] S. Dawson et al. Phys. Rev. D, 68(034022), 2003. [56] W. Beenakker et al. Nucl. Phys. B, 151(653), 2003. [57] M. Spira. Fortsch. Phys., 46(203), 1998. [58] The ATLAS Collaboration. Performance of the Reconstruction and Identification of Hadronic Tau Decays with ATLAS. Technical Report ATLAS-COM-CONF-2011-152, CERN, Geneva, Nov 2011. [59] Andreas Hoecker, Peter Speckmayer, Joerg Stelzer, Jan Therhaag, Eckhard von Toerne, and Helge Voss. TMVA: Toolkit for Multivariate Data Analysis. PoS, ACAT:040, 2007. [60] Rene Brun et al. Root website, 1995-2011. http://root.cern.ch/drupal/. [61] The ATLAS Collaboration. Identification of Hadronic Tau Decays for Summer 2011. Technical Report ATLAS-COM-CONF-2011-112, CERN, Geneva, Jul 2011. [62] The ATLAS Collaboration. Expected photon performance in the ATLAS experiment. ATL-PHYS-PUB-2011-007. [63] Data-Quality Requirements and Event Cleaning for Jets and Missing Transverse Energy Reconstruction with the ATLAS Detector in Proton-Proton Collisions at a Center-of-Mass Energy of sqrts=7 TeV. ATL-COM-PHYS-2010-247, ATLAS-CONF-2010-038. [64] ATLAS Collaboration. Phys. Rev. D, 84:112006, Dec 2011. [65] P. Nason S. Frixione and B.R. Webber. Matching NLO QCD and parton showers in heavy flavour production. JHEP, 0308:007, 2003. [66] G. Corcella et al. HERWIG 6.5: an event generator for Hadron Emission Reactions With Interfering Gluons (including supersymmetric processes). JHEP, 01:010, 2001. [67] J. R. Forshaw J. M. Butterworth and M. H. Seymour. Multiparton interactions in photoproduction at hera. Z. Phys, C72:637, 1996. [68] A. S. et al. Higgs boson production in gluon fusion. JHEP, 02:029, 2009. [69] J. H. Kuhn S. Jadach and Z. Was. Tauola - a library of monte carlo programs to simulate decays of polarized leptons. Comput. Phys. Commun., 64:275, 1990. [70] B. V. Eijk E. Barberio and Z. Was. Photos - a universal monte carlo for qed radiative corrections in decays. Comput. Phys. Commun., 66:115, 1991. [71] A. Elagin, P. Murat, A. Pranko, and A. Safonov. A New Mass Reconstruction Technique for Resonances Decaying to di-tau. Nucl.Inst.Meth., A654:481, 2011. [72] The ATLAS Collaboration. Luminosity Determination in pp Collisions at = 7 TeV using the ATLAS Detector in 2011. Technical Report ATLAS-COM-CONF-2011-130, CERN, Geneva, Jul 2011. [73] The ATLAS Collaboration, 2012. http://svn.cern.ch/reps/atlasoff/PhysicsAnalysis/MuonID/MuonIDAnalysis% /MuonEfficiencyCorrections/tags/MuonEfficiencyCorrections-01-00-10/. [74] The ATLAS Collaboration, 2012. http://svn.cern.ch/reps/atlasoff/Reconstruction/egamma/egammaAnalysis/ %tags/egammaAnalysisUtils-00-02-70/. [75] J Butterworth, E Dobson, U Klein, B Mellado Garcia, T Nunnemann, J Qian, D Rebuzzi, and R Tanaka. Single boson and diboson production cross sections in pp collisions at sqrts=7 tev. Technical Report ATL-COM-PHYS-2010-695, CERN, Geneva, Aug 2010. [76] J Butterworth, E Dobson, U Klein, B Mellado Garcia, T Nunnemann, J Qian, D Rebuzzi, and R Tanaka. Single boson and diboson production cross sections in pp collisions at sqrts=7 tev. Technical Report ATL-COM-PHYS-2010-695, CERN, Geneva, Aug 2010. [77] The ATLAS Collaboration. Close-by jet effects on jet energy scale calibration in pp collisions at sqrt(s)=7 tev with the atlas detector. Technical Report ATL-CONF-2011-062, CERN, 2011. [78] The ATLAS Collaboration, 2012. http://svn.cern.ch/reps/atlasoff/ Reconstruction/Jet/JetUncertainties/t%ags/JetUncertainties-00-03-04/. [79] The ATLAS Collaboration, 2012. http://svn.cern.ch/reps/atlasoff/ Reconstruction/MissingETUtility/tags/%MissingETUtility-00-02-12/. [80] K. Cranmer E. G. G. Cowan and O. Vitells. Asymptotic formulae for likelihood-based tests of new physics. Arxiv preprint hep-ex/10071727, 2010. [81] The ATLAS Collaboration. An update to the combined search for the standard model higgs boson with the atlas detector at the lhc using up to 4.9 fb1 of pp collision data at s = 7 tev. Technical Report ATLAS-CONF-2012-019, CERN, Geneva, Mar 2012.