Top quak mass measurement using m T2 at CDF (dilepton channel) - PowerPoint PPT Presentation

Top quak mass measurement using m T2 at CDF (dilepton channel) Hyunsu Lee The University of Chicago On behalf of the CDF collaboration PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Why we measure the top mass in the dilepton channel • It is important to check the mass crossing the channels � Is it SM top? � Significant difference indicate the new physics • This channel can be a standard candle for new physics search � Well known SM process � Signal and background is under control � Similar topology with new physics � Pair produced new particle can have two missing particle final state 2 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Signature of new physics particle and m T2 • New physics predict the candidate of dark matter (WIMP) � Ex) Neutralino in the SUSY • If we consider pair production of new physics particle � Two missing particle � Visible particle (quark and leptons) � Ex) two gluino pair production m T2 = m in[m ax(m T(1) ,m T(2) )] • We are interesting to determine the q T +p T =m issing p T mass of new particle � m T2 was introduced for two missing particle PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

m T2 • Transverse mass of two missing particle system � Similar with m T for W mass • Can be useful to determine the mass of new physics particle � One of the most stringent variable Tagged DIL Tagged DIL Arb units 0.08 • Top dilepton channel is good 0.07 2 M = 160 (GeV/c ) top example of m T2 variable (standard 0.06 2 M = 170 (GeV/c ) 0.05 top candle) 0.04 2 M = 180 (GeV/c ) top 0.03 • We can use real data 0.02 � First application in the real data 0.01 0 50 100 150 200 250 300 2 mT2 (GeV/c ) 4 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

ttbar dilepton channel • Dilepton (5% branching ratio, small background) 2 high-P T leptons(e/m), 2 b-jets, large missing ET � We separate subsample with b tagging PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Template method 2d templates MC reco + m T2 m t Event reconstruction � tt reco +H T m t � backgrounds 1d templates m T2 reco m t H T DATA Event reconstruction Likelihood Fit 6 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Sanity check Residuals: Dilepton mt+m Residuals: Dilepton mt+m Pull widths: Dilepton mt+m Pull widths: Dilepton mt+m T2 T2 T2 T2 2 1.2 ) pull width 2 Constant Constant 0.08424 0.08424 0.08797 0.08797 Constant Constant 1.006 1.006 0.005208 0.005208 residual (GeV/c ± ± ± ± 1.5 1.15 1 1.1 0.5 1.05 0 1 -0.5 0.95 -1 0.9 -1.5 -1 -1 CDF II preliminary 3.4 fb CDF II preliminary 3.4 fb -2 0.85 160 165 170 175 180 160 165 170 175 180 2 2 M (GeV/c ) M (GeV/c ) top top 2 Residuals: Dilepton mT2 only Residuals: Dilepton mT2 only 1.2 Pull widths: Dilepton mT2 only Pull widths: Dilepton mT2 only ) pull width 2 -0.2626 -0.2626 0.1047 0.1047 Constant Constant Constant Constant 0.9969 0.9969 0.005466 0.005466 residual (GeV/c ± ± ± ± 1.5 1.15 1 1.1 0.5 1.05 0 1 -0.5 0.95 -1 0.9 -1.5 -1 -1 CDF II preliminary 3.4 fb CDF II preliminary 3.4 fb -2 0.85 160 165 170 175 180 160 165 170 175 180 2 2 M (GeV/c ) M (GeV/c ) top top PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Expected statistical uncertainties 8 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Systematics and total estimated uncertainty 175 GeV/c 2 top mass assumed Unit (GeV/ c 2 ) m T2 give the best performance between single observables 9 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Data fit result m T2 and m t NWA m T2 alone 169.3 ± 2.7 (stat.) ± 3.2 GeV/c 2 (syst.) +4.8 (stat.) ± 2.9 GeV/c 2 (syst) 168.0 -4.0 10 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Data distribution NWA Non-tagged m t Non-tagged m T2 Non-tagged H T ) ) Events/(30 GeV/c) 2 2 30 Events/(10 GeV/c Events/(10 GeV/c 30 Data 25 Data Data 25 25 Signal+Bkgd Signal+Bkgd Signal+Bkgd 20 20 20 Bkgd only Bkgd only Bkgd only 15 15 15 10 10 10 -1 -1 CDF II Preliminary (3.4 fb ) CDF II Preliminary (3.4 fb ) -1 CDF II Preliminary (3.4 fb ) 5 5 5 0 0 0 100 150 200 250 300 350 50 100 150 200 250 300 200 300 400 500 600 700 800 2 2 2 m NWA (GeV/c ) H (GeV/c ) m (GeV/c ) T2 t t NWA tagged m t tagged m T2 tagged H T 24 ) ) Events/(30 GeV/c) 2 2 20 22 Events/(10 GeV/c Events/(10 GeV/c 22 18 Data Data 20 Data 20 18 16 18 Signal+Bkgd Signal+Bkgd Signal+Bkgd 16 14 16 14 12 14 Bkgd only Bkgd only Bkgd only 12 12 10 10 10 8 8 8 6 6 6 -1 -1 CDF II Preliminary (3.4 fb ) CDF II Preliminary (3.4 fb ) -1 4 CDF II Preliminary (3.4 fb ) 4 4 2 2 2 0 0 0 100 150 200 250 300 350 50 100 150 200 250 300 200 300 400 500 600 700 800 2 2 NWA H (GeV/c) m (GeV/c ) m (GeV/c ) t T2 t 11 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Conclusion • We measure the top quark mass in the dilepton channel using m T2 variable NWA and m T2 � 169.3 ±2.7 (stat.) ± 3.2 GeV/c 2 (syst.) with m t � 168.0 +4.8 (stat.) ± 2.9 GeV/c2(syst) with m T2 alone -4.0 � First application in the real data • We prove the performance of m T2 � Best single observable including systematic • This method can be useful to mass determination of new physics particle 12 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

m tNWA • Leptonical decay of top b � t->blv l � We measure b and lepton but W t don’t know neutrino ν � 4 unknown � Known parameter � W mass neutrino mass (2 unknown) � If we assume the top quark mass and neutrino eta direction, we can measure neutrino x,y momentum � Same thing happen for the other leg • Getting weight using measured missing transverse ν ν = η η w w m ( , , ) i i top 1 2 energy 13 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

m tNWA • Some over neutrino rapidities ∑∑ ∫ ∫ W ( m ) = d η d η P ( η ) P ( η ) w ( m ) t t i , j 1 2 1 2 j i • We have maximum weight m t NWA ) as reconstructed mass (m t • We scan mt with 3GeV size and then decrease the step size upto 0.15GeV near the peak • We have gaussian fit in the near of peak to get m t continuously 14 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

H t • Linear sum of jets, leptons, and missing Et • Strong JES correlation and also strong Top Mass NWA ) correlation (strong correlation with m t 15 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago

Why we measure top quark mass X ?? • SM Higgs Mass was constrained by M top and M W through loop correction of W mass • Precision top quark mass measurement � Predict SM Higgs mass � Constraints for physics beyond standard model 16 PHENO 2009 Symposium Hyunsu Lee, The University of Chicago