reconstruction Reconstruction strategy Reconstruct first Hadronic Top t

reconstruction § Reconstruction strategy § Reconstruct first Hadronic Top t jjb 2 & 1 b-tag : § Best procedure : reconstruct first W , See slide(s) 9 -11 § then choose b jet to form Had Top (ATL-com-phys 2008 -044) § Several equivalent methods to select the 2 jets from the W and the b jet from the Had TOP See slide(s) 11 No b-tag : § Define the jet triplet first => all recons meth could be used to cross-check the TOP reconstruction with first data => best standard methods have been studied in details => Other Recons strategy in development : Multi-Variate Analysis § Reconstruct then the Leptonic Top § => Etmiss + Mw constraint § ‘leptonic’ b jet from remaining jet D. Pallin Ringberg Wkshop 06/08 See slide(s) 22 See slide(s) 19 1

Ttbar Semileptonic channel Reconstruction 2 Btag W • • – – From AI Etienvre E Cogneras Reconstruction strategy Reconstruct first HADronic Top t jjb Best method : reconstruct first W , then choose B jet to form Had Top (ATL-com-phys 2008 -044) W reconstruction W jj 1. W 1 method : In each event, had. W defined as the light jet pair with lowest c² Þ reconstruction and in-situ rescaling in the same time • W Purity : 66. 6 0. 8 % W width : Gm. W=2. 1 Ge. V Constraint term Calibration coefficients 2. W 2 method : had. W defined as the light jet pair with lowest DR(jj) Þ very simple method, calibration done separately Þ independent of JES (relevant @ LHC start) • W Purity : 68. 8 % W width : m. W=10. 2 Ge. V Þ OR W defined as the light jet pair with lowest M(jj)-MW(peak) dependent of JES D. Pallin Ringberg Wkshop 06/08 2

Ttbar Semileptonic channel Reconstruction 2 Btag Had Top • From AI Etienvre E Cogneras b reconstruction t jjb, Choose the correct b-jet among 2 b-jets 1. In each event, b-jet defined as the Wb combination with lowest DR(Wb) Þ simplest method, calibration done separately Þ independent of JES (relevant @ LHC start) Þ W 1 meth Purity : 46. 5 % Top width : m. W=11. 4 Ge. V (W mass constraint) Þ W 2 meth Purity : 53. 8 % Top width : m. W =14. 3 Ge. V OR b-jet defined as the Wb combination with highest p. T(Wb) D. Pallin Ringberg Wkshop 06/08 3

Ttbar Semileptonic channel Reconstruction Nobtag, 1 Btag • – – – No b-tag reconstruction Same analysis as the 2 -btag choose jet triplet where the 3 jets are the closest in DR • • • Define the b-jet among the 3 jets The first light jet is the one with the lightest E in the jjb CM The second light jet is the closest jet to the 1 st one The b jet is the remaining jet • • 1 btag reconstruction to increase efficiency in top reconstruction No events in common with the 2 b-tag sample • Had Top reconstruction t jjb, • From E Cogneras W defined as the light jet pair with lowest M(jj)-MW(peak) – – Choose the correct b-jet among all remaining jets : the one compatible with a t jjb decay in the jjb centre-of-mass smallest (X 1 -mean 1) and (X 2 -mean 1) meanx 1 and meanx 2 extracted from data D. Pallin Ringberg Wkshop 06/08 4

Ttbar Semileptonic channel Reconstruction 0, 1 Btag • • – – – – From E Cogneras No B tag reconstruction Same analysis as the 2 -btag choose jet triplet where the 3 jets are the closest in DR • • • Define the b-jet among the 3 jets The first light jet is the one with the lightest E in the jjb CM The second light jet is the closest jet to the 1 st one The b jet is the remaing jet 1 btag reconstruction to increase efficiency in top reconstruction No events in common with the 2 b-tag sample W defined as the light jet pair with lowest M(jj)-MW(peak) Had Top reconstruction t jjb, Choose the correct b-jet among all remaining jets : the one compatible with a t jjb decay in the jjb centre-of-mass smallest (X 1 -mean 1) and (X 2 -mean 1) meanx 1 and meanx 2 extracted from data D. Pallin Ringberg Wkshop 06/08 5

Ttbar Semileptonic channel Purification cuts From AI Etienvre E Cogneras Apply purification cuts to remove fake jjb triplets (comb BKG) - Standard purification cuts Cut C 1 : Mass(jj) –MWpeak<2 m. W Remove mainly bad choice of jet pair to form the W Cut C 2 : Mass(lepton, W) >200 Ge. V. Cut C 3 : Mass(lepton, b-lept) <160 Ge. V. Cut C 6 : b-had closer to the W than to the lepton C 2 to C 6 Remove mainly bad choice of b-jet 2 b-tag 1 b-tag 0 btag - C 1+C 2+C 3 C 1+C 3+C 6 C 1+C 3 High purity purification cuts standard purification cuts + Cut C 4: |X 1−m 1| < 1. 5 1, Cut C 5: |X 2−m 2| < 2 2. D. Pallin Ringberg Wkshop 06/08 6

Purification cuts & mass measurement § Apply purification cuts to remove fake jjb triplets (comb BKG) See slide(s) 12 -15 § Standard purification cuts (purity up to 75%) § High purity purification cuts (purity up to 85%) => another possibility : don’t remove BKG but assess the BKG shape from data using an event See slide(s) 16 -18 mixing technique to reduce the syst from the BKG on the Mass measurement § Mass measurement 3 mass estimators § Mjjb See slide(s) 13 -15 § Mjjb – Mjj + Mw(peak) § Mtop from a Kinematical fit on the whole event § Uncertainties § Statistical & stability of the recons scheme and estimator : See slide(s) 20 See slide(s) 21 § evaluated using a bootstrap re-sampling Method § stat err ~0. 3 -0. 4 Ge. V @ 1 fb-1 § Systematics § mainly b jet scale (in-situ calib for light jets) & Bkg shape (no-b tag) §pile-up ? Jet scale dependence %E, ? Bias from trigger? Other bias? D. Pallin Ringberg Wkshop 06/08 7

Ttbar Semileptonic channel Top mass spectra 2 btag - Standard purification cuts From AI Etienvre E Cogneras High purification cuts METH W 2 METH W 1 D. Pallin Ringberg Wkshop 06/08 8

Ttbar Semileptonic channel Top mass spectra 1 b tag - Standard purification cuts From E Cogneras High purification cuts D. Pallin Ringberg Wkshop 06/08 9

Ttbar Semileptonic channel Top mass spectra 0 b tag - Standard purification cuts From E Cogneras High purification cuts D. Pallin Ringberg Wkshop 06/08 10

Ttbar Semileptonic channel Top mass measurement uncerntainty From AI Etienvre E Cogneras Statistical uncertainty & Stability of the methods evaluated using a bootstrap resampling technique Produced the pull distributions for 1200 pseudo-experiments. => No biais, ~03 -0. 4 Ge. V depending of the used method Systematics Light jet scale dependance is small B-jet scale is the major contribution to the systématics BKG shape syst could be important for 1 b-tag and 0 b –tag , but reduced largely by the use of an event mixing technique Pile-up ? D. Pallin Ringberg Wkshop 06/08 11