Top-quark properties and mass measurements with the ATLAS detector

The top quark is unique among the known quarks in that it decays before it has an opportunity to form hadronic bound states. This makes measurements of its properties particularly interesting as one can access directly the properties of a bare quark. The latest measurements of these properties with the ATLAS detector at the LHC [1] are presented. Measurements of top-quark spin observables in top-antitop events, each sensitive to a different coefficient of the spin density matrix, are presented and compared to the Standard Model predictions. The helicity of the W boson from the top decays and the production angles of the top quark are further discussed. Limits on the rate of flavour changing neutral currents in the production or decay of the top quark are reported. The production of top-quark pairs in association with W and Z bosons is also presented. The measurement probes the coupling between the top quark and the Z boson. The cross-section measurement of photons produced in association with top-quark pairs is also discussed. These process are all compared to the best available theoretical calculations. The latest ATLAS measurements of the top-quark mass in lepton+jets, dilepton, and all-hadronic final states are also reported. In addition, measurements aiming to measure the mass in a well-defined scheme are presented.


Introduction
With the mass around 173 GeV , the top quark is the heaviest known elementary particle. Due to its large mass, the top quark in the Standard Model (SM) has large decay width corresponding to very short mean life-time of around 0.5 × 10 −24 s. Because of this short mean life-time, the top quark decays before it can form a bound state and transfers all its properties to its decay products. This makes the top quark unique among other quarks, as it allows to study properties of a bare quark. Moreover, the top quark Yukawa coupling to Higgs boson is O(1) which implies that the top quark could play a special role in the process of electroweak symmetry breaking. Many Beyond Standard Model (BSM) scenarios predict different values for some of the properties of the top quark. Thus precisely measuring the properties of the top quark can test SM predictions and constrain some of the BSM parameters. In addition, the top quark is produced at very short distances -a characteristic strong coupling constant α s is of the order of 0.1. This makes top quark a perfect object to study perturbative QCD. a e-mail: tomas.dado@cern.ch EPJ Web of Conferences 182, 02033 (2018) https://doi.org/10.1051/epjconf/201818202033 ICNFP 2017 Currently, the Large Hadron Collider (LHC) is referred as "a top factory", since the top quarks are produced abundantly during collisions. This makes the LHC an ideal place to measure the topquark properties. Top quarks are predominantly produced in pairs in the LHC, however single-top production is also possible via electroweak production mechanism. In the SM the top quark decays almost exclusively into W boson and b-quark. Depending on the subsequent decay of the W boson there are three decay channels of top-quark pair: all-hadronic -both W bosons decay into pair of quarks, semi-leptonic -one W boson decays into pair of quarks and the other W boson decays into charged lepton and its corresponding neutrino and finally dileptonic -both W bosons decay into charged leptons and their corresponding neutrino.

Measurement of top-quark spin observables
Top quarks are assumed to be produced unpolarised, which means their spins are not aligned with any direction. However, spins of top and antitop quark from pair production are correlated. The level of correlation depends on reference quantisation axis and the production process. Top quarks decay before any spin-flip occurs, thus the full information about the quark spin is transferred to its decay products.
The most recent ATLAS result [2] analyses fifteen observables sensitive to top-quark pairs' spin density matrix. The measurement uses tt dileptonic data at centre-of-mass energy √ s = 8 TeV. Neutrino weighting [3,4] technique is used to fully reconstruct tt kinematics. Normalised doubledifferential cross-section for tt production and decay is of the form 1 σ where B a , B b and C(a, b) are the polarisation and spin correlations along the spin quantisation axes, a and b. Angles θ a and θ b are defined as angles between the momentum direction of a top-quark decay particle in its parent top-quark's rest frame and the axis a or b respectively. Three different quantization axes are compared: helicity axis (k) -defined as the top-quark direction in the tt rest frame, transverse axis (n) -defined to be transverse to the production plane created by top-quark direction and the beam axis and r-axis (r) -axis orthogonal to the other two axes. Detector level distributions are unfolded to parton level, shown in Figure 1 and compared to theory predictions. Results are also unfolded to stable particle level and are compared to prediction from MC simulation (Powheg-hvq+Pythia6). At parton level, the measurements along the helicity axes are

W helicity measurement in top-quark events
In SM, top quark decays almost exclusively to W boson and b quark. The Wtb decay vertex has V − A structure, where V and A refer to the vector and axial vector components of the weak coupling. Due to this structure and large difference between top quark and b quark mass, the W boson from top quark decay has only left-handed and longitudinal polarisation, the right-handed polarisation is heavily suppressed. The W helicity in semileptonic decays of top-quark pair is measured with the ATLAS detector at centre-of-mass energy √ s = 8 TeV [5]. An observable sensitive to the W boson polarisation is angle θ * -angle between b-quark from the top-quark decay and a direct W boson decay product in the W boson rest frame. Polarisations of leptonically and hadronically decaying W bosons are analysed. Charged lepton and down-type quark are used as direct decay products of W bosons in leptonically and hadronically decaying channels respectively. Extended KLFitter [7] technique is used to fully reconstruct tt kinematics. The extension enables separation between up-and down-type quark from the W decay. Pure helicity templates of cos θ * distributions are fitted to data. The measured W helicities as well as theoretical NNLO predictions [8] are summarized in Table 1. Dominant systematic uncertainties arise from uncertainties in jet energy reconstruction and signal modelling. No significant deviation from SM prediction is observed, thus the limits on possible anomalous couplings in Effective Field Theory, e.g. (EFT) [6], are calculated and presented in Figure 2. This is the most precise W helicity measurement till date.

Direct measurement of the top-quark decay width
Decay width is an important property of every elementary particle. In SM, the decay width depends on the particle mass. NNLO prediction [9] yields top-quark decay width Γ t = 1.32 GeV for the top-quark mass m t = 172.5 GeV.
The first direct measurement with the ATLAS detector [10] uses 20.2 fb −1 of data at centre-of-mass energy √ s = 8 TeV. The measurement focuses on semileptonic decay of tt. Two variables sensitive to decay width of top quark are used: m b -invariant mass of lepton and corresponding b-jet from top-quark decay and ∆R min ( j b , j light ) -∆R † between b-jet from hadronically decaying top and closest light jet from hadronically decaying W boson. Events are split by lepton flavour (e or µ), number of b-tagged jets (exactly one b-tagged jet and at least two b-tagged jets) and jet |η| = 1 to decrease systematic uncertainties in jet energy reconstruction. Templates with different widths are fitted to data and the measured width is assuming top mass m t = 172.5 GeV. Measurement is limited by jet energy reconstruction and signal modelling uncertainties. The measured width is in good agreement with the SM prediction.

Searches for FCNC with top quarks
Flavour-changing neutral currents (FCNCs) are forbidden at tree level in the SM and are heavily suppressed by the GIM mechanism. However, different BSM scenarios predict large effective couplings which are orders of magnitude larger than those in the SM. Searches for FCNC can give either a hint or constrain some of the BSM scenarios.

Search for t → qH(H → γγ)
The ATLAS collaboration performed a search [11] for t → qH(→ γγ), where q = c, u; at the centreof-mass energy √ s = 13 TeV, corresponding to 36.1 fb −1 . In SM, this process is heavily suppressed with branching ratio of approximately 3 × 10 −15 . Two event selections are applied: hadronic selection where the non-FCNC top quark decays hadronically t → bW(W → j j) and leptonic selection where t → bW(W → ν). Events are required to have diphoton mass 100 < m γγ < 160 GeV. Events are further split into two complementary categories. In the first category events passing all selection criteria for hadronic or leptonic selection are considered. In the second category events passing all selection criteria except the criterion on reconstructed top-quark mass 120 < m j j j < 220 GeV for hadronic selection and 130 < m j ν < 210 GeV for leptonic selection are considered. Control regions are used to constrain dominant backgrounds: associated production of top-quark pair with a vector boson and diboson processes. No significant excess of events over expectation is observed. Limits on FCNC branching ratios are set This corresponds to limits on anomalous couplings assuming m H = 125.09 GeV. Dominant systematic uncertainties arise from jet energy scale reconstruction and signal modelling. Current sensitivity allows for a probe in regions of flavour violating Yukawa coupling in Two Higgs Doublet Models (2HDM).

Search for t → qZ(Z → )
A search for t → qZ(Z → ) is performed by the ATLAS collaboration using centre-of-mass √ s = 8 TeV data, corresponding to 20.3 fb −1 [12]. Branching ratio predicted by SM for this process is of the order of 10 −14 . The search focuses on top-quark pair events with Z boson decaying leptonically (Z → ) and the non-FCNC top quark decaying semileptonically t → bW(W → ν). Kinematics of tt system is reconstructed using χ 2 kinematic method. Additionally, χ 2 < 6 is required to improve the purity of the selection. No significant deviation from SM prediction is observed and the limits on t → qZ are set as Uncertainty in the background modelling is the dominant systematic uncertainty of this measurement. Combination with other FCNC searchers is shown in Figure 3.

Search for qg → t
Searching for FCNC decays of top quark to an up-type quark and gluon (t → qg) is extremely difficult due to overwhelming QCD background. Alternatively, one may look for FCNC production of single top quark. This search is performed by the ATLAS collaboration [13] using centre-of-mass energy  Figure 3. The current 95% CL observed limits on the (a) BR(t → qγ) vs BR(t → qZ) and (b) BR(t → qH) vs BR(t → qZ) planes are shown. The ATLAS lines correspond to the limit on BR(t → qZ) set in this measurement [12].
properties: FCNC single top-quark p T distribution is softer compared to single top SM prediction; the W boson from the top-quark decay has high momentum and its decay products tend to have small angles and the top-quark charge asymmetry differs between FCNC processes and SM processes in the ugt channel. Unfortunately, none of the observables have enough separation power to distinguish top quarks produced via FCNC. Thus artificial neural network (NN) is used. NN response is fitted to observed data. Control region is used to constrain the dominant W+jets background. No significant excess over SM prediction is observed and limit on production cross section is set as σ qg→t × B(t → Wb) < 3.4 pb at 95% CL.
The limit on production cross section is translated to limit on coupling constant divided by scale factor of new physics (Λ) and on branching fractions (B)   all other processes are zero Each limit assumes that Figure 5. Summary of the current 95% confidence level observed limits on the branching ratios of the top-quark decays via flavour changing neutral currents to a quark and a neutral boson t → Xq(X = g, Z, γ or H; q = u or c) by the ATLAS and CMS Collaborations compared to several new physics models [14].
cross sections from data collected in 2015 using centre-of-mass energy √ s = 13 TeV, corresponding to integrated luminosity of 3.2 fb −1 [15]. The measurement focuses on three different signal regions: same-sign di-muon channel -targeting ttW production; trilepton channel -targeting both ttW and ttZ production and finally tetralepton channel -targeting ttZ production. Each analysis channel is further divided into multiple regions in order to enhance the sensitivity to signal. In total, nine signal and two control regions enter the profile likelihood fit in order to extract the cross sections for ttW and ttZ production. Only semileptonic and dileptonic decay channels of top-quark pair are considered. with the observed significance of 3.9 σ and 2.2 σ with respect to background-only hypothesis for ttZ and ttW respectively. The correlation between the observed cross sections is 13%. Figure 6 shows observed cross sections for ttW and ttZ with their uncertainties. The obtained results are in good agreement with the SM predictions.

Associated production of tt and photon (γ)
Total and differential production cross section of tt + γ is measured with the ATLAS detector using centre-of-mass energy √ s = 8 TeV [16]. The measurement focuses on semileptonic decay of the tt system. The events are separated into three categories: a) events with prompt photon originating from ttγ matrix element, b) events with photon from hadron decays or hadrons misidentified as photons and c) events with electrons misidentified as photons. A likelihood fit for the templates in the three categories is used to extract the total and differential cross sections of the process. The cross sections are measured within fiducial volume close to the selection requirements in the measurement and compared to theoretical NLO predictions [17]. The measured cross section is The measurement is limited by the uncertainty of the photon fakes modelling. Differential cross section as a function of photon |η| and photon p T is also measured and no significant deviation from SM prediction is observed.

Dilepton channel
The most precise measurements of m top exploit MC generators to create templates with different assumptions on m top . These templates are then fitted to data and the top-quark mass is extracted.
The ATLAS collaboration measures m top using dilepton tt events at centre-of-mass energy √ s = 8 TeV [18], corresponding to integrated luminosity of 20.

All-hadronic channel
The ATLAS collaboration measures top-quark mass in all-hadronic tt channel [20]. The measurement benefits from the largest branching ratio (46%) among the possible top-quark decay channels. However, the measurement is difficult due to large multi-jet background, which can exceed the tt production by several orders of magnitude. To reduce the overwhelming multi-jet background, events are required to have at least 5 jets with tight selection of p T > 60 GeV. The two jets with largest b-tag weight are required to have ∆φ(b i , b j ) > 1.5. This requirement is very powerful in rejecting combinatorial background. Kinematics of tt system is reconstructed using χ 2 method. The observable sensitive to m top is ratio of three-jet mass divided by two-jet mass (R 3/2 ). Templates of R 3/2 are generated for discrete m top values. The templates are parametrized by combination of a Novosibirsk and a Landau function. Multi-jet background is estimated using ABCD method. Statistical uncertainty is corrected by a factor of 1 + ρ due to the correlation of two with relative uncertainty of 0.7%. Dominant systematic uncertainties arise from jet energy reconstruction and hadronization modelling. Figure 7 shows post-fit distribution of R 3/2 and correlation of measured m top and fitted background normalization. Summary of recent ATLAS m top measurements is shown in Figure 8.

Top-quark pole mass measurement using lepton distributions
Very recent measurement by ATLAS collaboration exploits lepton distributions in the eµ channel of tt events at centre-of-mass energy √ s = 8 TeV [23]. Some of the differential cross sections of leptonic observables are sensitive to top-quark mass. Measured distributions are compared to prediction from both NLO plus parton shower event generators and fixed-order QCD calculations. The former are similar to traditional MC mass measurements but are less sensitive to hadronization modelling uncertainties as they rely purely on leptonic observables. Measurements based on fixed-order QCD calculations in a well-defined renormalization scheme correspond more directly to m

Conclusions
This proceedings contribution reviewed some of the most recent results on the top-quark properties measurements performed by the ATLAS detector. In particular measurements of the top-quark spin correlations, W helicity, top-quark decay width, top-quark mass measurements, including top-quark pole mass measurements and cross sections of associated production of tt and W/Z boson or photon are reported. Additionally, searches for FCNC involving top quarks are summarized.