t b Measurement of t b cross-section in 13 TeV CMS data and - - PowerPoint PPT Presentation

Measurement of t¯ t → b¯ bℓτ 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 100fb−1

• f 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

Cross-section measurements in t¯ t channels at 8 TeV CMS data from 2012.

Measurement in t¯ t → b¯ blτ 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

Oleksii Toldaiev (supervised by Michele Galinaro and Joao VarelaLIP, CMS) t¯ t → ℓτ cross-section 16 February 2018 3 / 11

Features of t¯ t → b¯ bℓτ 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 → ℓνℓq¯ q channel. The events are separated into background-rich and signal-rich categories according to kinematics of jets. The shape fit of MT(ℓ, E miss

T

) distributions is performed. 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 ℓj provides separation between this background and signal via the shape of transverse mass distributions MT(ℓ, E miss

T

) and kinematic difference in jets.

transverse mass MT(ℓ, E miss

T

)

20 40 60 80 100 120 140 160 180 200 50 100 150 200 250 300 50 100 150 200 250 300 20 40 60 80 100 120 140 160 180 200 50 100 150 200 250 300 50 100 150 200 250 300

masses of jet combinations for W and t mass 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 MT distribution. Likelihood function includes per-bin yields and systematic uncertainties as constraint nuisance parameters: L(µ, θi) =

• k

Poisson

• Nk| ˆ

Nk(µ, θi)

• ·
• i

pdf (θi, 0, 1) (1) Based on the likelihood function the PLR test statistic is defined: λ(µ) = L(µ, ˆ ˆ θi(µ)) L(ˆ µ, ˆ θi) (2) — 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 eτh and µτh

Scan of signal strength.

16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1

θ ∆ )/ θ

• θ

(

2 − 1 − 1 2

JER JES ISR HDAMP qcd_norm PU wjets_norm TuneCUETP8M2T4 TauES tau_fakes bSF FSR TOPPT lumi_13TeV dy_norm tauID_eff

0.277 − 0.264 +

0.222

CMS Internal

r ∆

0.05 − 0.05 Pull Impact σ +1 Impact σ

• 1

Unconstrained Gaussian Poisson AsymmetricGaussian

0.0673 ± = 0.989 r

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

The goal is to measure precisely (on order 2%) the ratio

W →ℓν W →τν :

+

W

• µ

t g

+

τ

τ

ν g t b b

• W

µ

ν

+

τ

+

π

+

π

• π

+

W

• µ

t g

+

µ

τ

ν g t b b

• W

µ

ν

The ratio cancels most of systematic uncertainties: σ(µτ) = σpp(t¯ t)B(W → µ)B(W → τ) σ(µµ) = σpp(t¯ t)B(W → µ)(B(W → µ) + B(W → τ → µ)) (3) σ(µτ) σ(µµ) = B(W → τ) B(W → µ) + B(W → τ → µ) =

B(W →τ) B(W →µ)

1 + B(W →τ)

B(W →µ)B(τ → µ)

(4) 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 2.5σ 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

• ne 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

• ther 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