t → b ¯ Measurement of t ¯ b ℓτ cross-section in 13 TeV CMS data and lepton universality test Oleksii Toldaiev supervised by Michele Galinaro and Joao Varela LIP, CMS 16 February 2018 Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 1 / 11
CMS detector, available data The LHC schedule and luminosity collected by CMS, promising perspective of 100 fb − 1 of 13 TeV data at end of Run 2. Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 2 / 11
Motivation t → b ¯ Measurement in t ¯ bl τ channel Improved uncertainty in estimation of main background It serves as preliminary work for further measurements in similar final states The plan is to proceed to precise lepton universality measurement in t ¯ t decay Cross-section measurements in t ¯ t channels at 8 TeV CMS data from 2012. Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 3 / 11
t → b ¯ Features of t ¯ b ℓτ channel, measurement method Many particular final products: 2 b-jets (displaced vertex of jet) 1 lepton (muon or electron) 2 neutrinos 1 hadronicaly decaying tau lepton Sample of t ¯ t events is selected with simple cuts and identification requirements. Main background from fake taus in t ¯ t → ℓν ℓ q ¯ q channel. The events are separated into background-rich and signal-rich categories according to kinematics of jets. The shape fit of M T ( ℓ, E miss ) distributions is performed. T Both methods constrain background of misidentified taus and cross-check each other. Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 4 / 11
Reconstruction algorithms and event selection Standard algorithms are employed: Particle Flow for basic objects, anti-Kt jet clustering, MVA-based b-tagging, quality requirements for muons and electrons, MVA-based tau ID etc. Require: 1 lepton, ≥ 3 jets, ≥ 1 b-tagged and 1 tau lepton. no tau requirement tau of Opposite Sign to muon Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 5 / 11
Background of misidentified taus No loss of kinematic information in W → q ¯ q of the background ℓ j provides separation between this background and signal via the shape of transverse mass distributions M T ( ℓ, E miss ) and kinematic difference in jets. T 300 300 300 300 250 250 250 250 200 200 200 200 150 150 150 150 100 100 100 100 50 50 50 50 0 0 0 0 0 20 40 60 80 100 120 140 160 180 200 0 20 40 60 80 100 120 140 160 180 200 transverse mass masses of jet combinations for W and t mass M T ( ℓ, E miss ) T constraint in signal (left) and background (right) Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 6 / 11
Profile Likelihood Ratio (PLR) shape fit in two categories Background- or signal-rich categories are defined by jet kinematic parameter. Profile likelihood ratio fit is performed in bins of M T distribution. Likelihood function includes per-bin yields and systematic uncertainties as constraint nuisance parameters: � � � N k | ˆ � L ( µ , θ i ) = P oisson N k ( µ , θ i ) · pdf ( θ i , 0 , 1) (1) k i Based on the likelihood function the PLR test statistic is defined: λ ( µ ) = L ( µ , ˆ ˆ θ i ( µ )) (2) µ , ˆ L (ˆ θ i ) — scans over λ ( µ ) provide estimation of uncertainties. Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 7 / 11
Preliminary results for fit in both e τ h and µτ h Unconstrained Gaussian CMS Internal ± r = 0.989 0.0673 Poisson AsymmetricGaussian tauID_eff 1 dy_norm 2 lumi_13TeV 3 TOPPT 4 FSR 5 bSF 6 + tau_fakes 0.222 0.264 7 − 0.277 TauES 8 TuneCUETP8M2T4 9 wjets_norm 10 PU 11 qcd_norm 12 13 HDAMP 14 ISR Scan of signal strength. 15 JES JER 16 − − − 2 1 0 1 2 0.05 0 0.05 θ θ ∆ θ ∆ σ σ ( - )/ r Pull +1 Impact -1 Impact 0 Impacts of uncertainties on signal strength. Results show agreement with SM and uncertainty of about 6-7% in both channels. Plots show simultaneous fit over both channels. The largest uncertainty is 5% from Tau ID. Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 8 / 11
Prospects, lepton universality test W → ℓν The goal is to measure precisely (on order 2%) the ratio W → τν : + + + µ τ π + + W + W τ + π - π g ν g ν τ τ t t b b b b t - t - µ µ g g - - W W ν ν µ µ The ratio cancels most of systematic uncertainties: σ ( µτ ) = σ pp ( t ¯ t ) B ( W → µ ) B ( W → τ ) (3) σ ( µµ ) = σ pp ( t ¯ t ) B ( W → µ )( B ( W → µ ) + B ( W → τ → µ )) B ( W → τ ) σ ( µτ ) B ( W → τ ) B ( W → µ ) σ ( µµ ) = B ( W → µ ) + B ( W → τ → µ ) = (4) 1 + B ( W → τ ) B ( W → µ ) B ( τ → µ ) But the remaining uncertainty due to tau ID is big (about 5%). Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 9 / 11
Current measurement from Particle Data Group (2012) measurements from LEP in WW channel Tevatron in W+jets excess of about 2 . 5 σ with relative uncertainty ≈ 3 . 5% at LHC: enough energy for on-shell t ¯ t and a lot of luminosity current measurements lack precision (about 6-10%, when 2% needed) with the luminosity we can sacrifice efficiency for purity Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 10 / 11
Plans Finalizing the measurement of the top quark section in the final state with one tau lepton On-going investigation of possibilities to improve tau ID for the ratio measurement includes: simultaneous fit with DY processes tau parameters: Secondary Vertex, Dalitz parameters of the decay other physics in the event: better b-tagging, kinematics and OS/SS contribution of backgrounds and machine learning algorithms based on these inputs Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t ¯ t → ℓτ cross-section 16 February 2018 11 / 11
Recommend
More recommend
