1. Top Physics ATLAS overview Stan Bentvelsen ATLAS-UK Physics meeting, IPPP Durham 18-20 September 2006 From early data to precision measurements

2. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 2 5 orders of magnitude Motivations for top quark physics  Special role in the EW sector and in QCD  Heaviest elementary particle known  Yukawa coupling close to 1.0  Top and W masses constrain the Higgs mass  A tool to probe symmetry breaking in SM  Special role in various SM extensions through EWSB  New physics often preferentially coupled to top  New particles can produce / decay to tops  A sensitive probe to new physics  Special interest even if it is just a «normal» quark  A major source of background for many searches  A tool to understand/calibrate the detector, all sub-detectors involved

3. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 3 Top properties scorecard We still know little about the top quark, limited by Tevatron statistics Mass precision <2% Electric charge ⅔ -4/3 excluded @ 94% C.L. (preliminary) Spin ½not really tested – spin correlations BR to b quark ~ 100%at 20% level in 3 generations case V – A decayat 20% level FCNCprobed at the 10% level Top width?? First observe single top ! Yukawa coupling??  This leaves plenty of room for new physics in top production and decay  Tevatron run II starts to incise probe the top quark sector  The LHC will open a new opportunity for precision measurements

4. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 4 10% 90% Production: σ tt (LHC) ~ 830 ± 100 pb  1 tt-event per second Top quark production at the LHC Cross section LHC = 100 x Tevatron Background LHC= 10 x Tevatron t t Final states: t  Wb ~ 1 W  qq ~ 2/3 W  l ν ~ 1/3 1) Full hadronic (4/9) 6 jets 2) Semi-leptonic (4/9): 1l + 1 ν + 4 jets 3) Full leptonic (1/9): 2l + 2 ν + 2 jets Golden channel (l=e, μ )  2.5 million events/year

5. Earlier studies on Atlfast: Selected 87000 signal events for L=10 fb -1 (S/B~78) now being repeated by full simulation Systematics uncertainties: Jet energy calibration ‘Out of cone’ showering B-fragmentation (  =3.5%) Precision top mass (physics TDR) Source of uncertainty Hadronic  M top (GeV) Fitted  M top (GeV) Light jet scale 0.90.2 b-jet scale0.7 b-quark fragm 0.1 ISR0.1 FSR1.90.5 Comb bkg0.40.1 Total2.30.9 The challenge is to determine m top below 1 GeV level (with L=10fb-1) Comb.  =10.6GeV

6. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 6 Alternative mass determination Select high P T back-to-back top events:  Hemisphere separation (bckgnd reduction, much less combinatorial)  Higher probability for jet overlapping Use the events where both W’s decay leptonically (Br~5%)  Much cleaner environment  Less information available from two ’s Use events where both W’s decay hadronically (Br~45%)  Difficult ‘jet’ environment  Select P T >200 GeV Other ideas:  Determine M top from J/Y production  Use the decay length of B-hadron (see Dimitri Typaldos) M top Various methods all have different systematics

7. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 7 Consequences from M top Assuming total uncertainty on W-mass of 15 MeV  Combined LHC prospect  Very challenging measurement! Repeat the Electro-Weak fit changing the uncertainties:   M top =1 GeV   M W =15 MeV  Same central values Summer 2005 values: Valuable check of SM! (  m H /m H  53%) Thanks to M.Grunewald direct EXCLUDED Increase the reality content on the top mass determination…

8. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 8 List of top event generators Leading Order MC:  Pythia & Herwig : full standalone MC  TopRex (include spin effects – interfaced to Pythia)  AcerMC (include spin effects – interfaced to Pythia)  AlpGen (include additional hard jets)  Sherpa (full MC, needs validation…) NLO QCD calculations implemented in MC  MC@NLO – interfaced to Herwig shower and fragmentation  ttbar process (among others) available  Recently single top processes included (s- and t-channel)  This is theoretical improvement  Validation done for this generator  Widely used in Atlas now Events have ‘negative weight’ 86.5%13.5%

9. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 9 Generators: MC@NLO, Herwig, Pythia Examples of MC validation (fast simulation detector response) PT(tt system) PT system balanced by ISR & FSR Herwig & MCatNLO agree at low PT, At large PT MCatNLO ‘harder’ PYTHIA completely off Jet PT Jet with highest PT Top mass Standard reconstruction using 1 b-tagged jet Geant4 vs fast simulation G4 fast

10. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 10 AcerMC versus MC@NLO (G4 sim) Distributions not compatible  Fit (gaussian + P3)  4 GeV difference !! Ongoing work MC@NLO AcerMC MC@NLOAcerMC <>169.2 ± 0.4165.0 ± 0.6  10.5 ± 0.412.7 ± 0.7 Number of add jets with P T > 10 GeV MC@NLO AcerMC Number of extra jets Reconstructed mass

11. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 11 UE / min-bias determination Extremely difficult to predict the magnitude of the UE at LHC  Will have to learn much more from Tevatron before startup  Energy dependence of dN/d  ? High precision top mass only after well understood detector & underlying physics LHC? at  = 0 M top for various UE models Top peak for standard analysis using 2 b-tags Difference can be as large as 5 GeV

12. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 12 Top analysis w/o b-tag First apply selection cuts Assign jets to W, top decays 1 lepton P T > 20 GeV Missing E T > 20 GeV 4 jets(R=0.4) P T > 40 GeV Selection efficiency = 5.3% TOP CANDIDATE 1 Hadronic top: Three jets with highest vector-sum pT as the decay products of the top 2 W boson: Two jets in hadronic top with highest momentum in reconstructed jjj C.M. frame. W CANDIDATE

13. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 13 Signal-only distributions M W = 78.1±0.8 GeV m top = 162.7±0.8 GeV S/B = 1.20 S/B = 0.5 S B m(top had )m(W had ) TOP CANDIDAT E W CANDIDATE Clear top, W mass peaks visible Background due to mis-assignment of jets  Easier to get top assignment right than to get W assignment right Masses shifted somewhat low  Effect of (imperfect) energy calibration Jet energy scale calibration possible from shift in m(W) L=300 pb -1

14. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 14 Estimate W+jets bckgrnd Sources of physics background  W+njet production Only using W + 4 partons now, but W + 3,5 partons may also result in W + 4 jet final state due to splitting/merging MLM matching prescription to explicit elimination of double counting.  W+bbar, WW, ZZ, etc W  l W + 4 partons (32 pb * ) W + 3 partons (80 pb * ) W + 5 partons (15 pb * ) parton is reconstructed as 2 jets 2 parton reconstructed as single jets AlpGen v2.03: in progress Convincing at Tevatron e.g. PT spectrum W * These are the cross sections with the analysis cuts on lepton and jet pT applied at the truth level This background cannot be simulated by Pythia or Herwig shower process. Dedicated generator needed: e.g. AlpGen. Large uncertainties in rate

15. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 15 MLM matched W+jets background simulations Ongoing work:  Distribution of jet p T s with breakdown by sample  Histograms show cumulative distributions of W+0,1,2,3,4,5 parton samples  HUGE amounts of CPU needed for these studies  Conclusions are very similar to using W+4jets with sensible settings I.e. “A7 Rome sample”  Need to produce the W+bjets background… pT(jet1)pT(jet2) pT(match)=40 pT(match)=20

16. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 16 Estimating QCD background QCD background  Not possible to realistically generate Crucially depends on Atlas’ capabilities to separate  /e to 10 -5  Background has to be obtained from data itself Crude estimate using jets in ttbar events, O(10 5 ) jets in CSC samples  Determine the isolated electron fake rate Estimated jet  isolated electron fake rates (T1 sample)  Light jets: 9 / ( 2  28400) = 1.6 x 10 -4  b jets: 15 / (2  28400) = 2.6 x 10 -4 Not bad, but not quite 10 -5 … e-,0e-,0 PDG-id of patched parton b quarks light quarks (very preliminary!) “fake electron” is required to lign up with the underlying parton Opening angle parton-electron

17. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 17 Signal + Wjets background S/B = 0.45 S/B = 0.27 S B m(top had )m(W had ) TOP CANDIDAT E W CANDIDATE Plots now include W+jets background  Background level roughly triples  Signal still well visible  Caveat: bkg. cross section quite uncertain Jet energy scale calibration possible from shift in m(W) L=300 pb -1 (~1 week of running)

18. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 18 Results for 100-200 pb -1 (with cut on W mass) M jjj (GeV) electron-only L =100 pb -1 electron+muon estimate for L=100 pb -1 L=200 pb -1 Hadronic 3-jet mass Events / 4.15 GeV Hadronic 3-jet mass Distribution of 3-jet invariant mass after a cut on the mass of the reconstructed W-boson: 70 < Mjj < 90 GeV

19. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 19 Exploiting ttbar for JES calibration Use the W mass constraint to set the JES  Rescale jet E and angles to parton energy  = E parton / E jet  Takes into account variation of rescaling parameter with energy and correlation between energies and opening angle before after P T cut = 40 GeV All jj combinations Only 2 light jets + 150 < m jjb < 200 Only 2 light jets In cooperation with Jet/ETMiss  Newly developed ‘template’ method

20. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 20 Exploiting ttbar as b-jet sample TOP CANDIDAT E W CANDIDATE Simple demonstration use of ttbar sample to provide b enriched jet sample  Cut on m(W had ) and m(top had ) masses  Look at b-jet prob for 4 th jet (must be b- jet if all assignments are correct) W+jets (background) ‘random jet’, no b enhancement expected AOD b-jet probability Clear enhancement observed! ttbar (signal) ‘always b jet if all jet assignment are OK’ b enrichment expected and observed

21. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 21 Use b-tag of 4 th jet ttbar W+jets Standard analysis (for comparison) Cut on b signal probability > 0.90 on 4th jet ttbar W+jets Use b-tag of 4th jet to clean up hadronic top S/B = 0.45S/B = 1.36 Effectively background is killed, combinatorics left Backgnd*2

22. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 22 Effect of trigger – 1 st steps Study started with e/γ trigger Triggering through 2E15i, E25i, E60 channels  Preliminary trigger efficiency as function of lepton p T Includes effects of ‘untriggerable’ events due to cracks etc… Work in progress; important topic for CSC samples Electron pT (GeV) Reco elec Reco+trig elec E/γ trig eff (pT)E/γ trig eff (η) Preliminary! Trigger for top in next talk by Simon Head

23. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 23 Top physics during commissioning Inputs Single lepton trigger efficiency Lepton identification efficiency Integrated luminosity At startup around 10-20%. Ultimate precision < 5% What we can provide Top enriched samples Estimate of a light jet energy scale Estimate of the b-tagging efficiency Calibration of missing energy Estimate of M top and σ top ~20% accuracy. One of ATLAS’ first physics measurements? Can reconstruct top and W signal after ~ one week of data taking without using b tagging

24. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 24 Top charge determination Can we establish Q top =2/3?  Currently cannot exclude exotic possibility Q top =-4/3 Assign the ‘wrong’ W to the b-quark in top decays  t  W - b with Q top =-4/3 instead of t  W + b with Q top =2/3 ? Technique:  Hard  radiation from top quarks Radiative top production, pp  tt  cross section proportional to Q 2 top Radiative top decay, t  Wb   On-mass approach for decaying top: two processes treated independently Matrix elements have been calculated and fed into Pythia MC Radiative top production Radiative top decay

25. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 25 Top Charge determination Yield of radiative photons allows to distinguish top charge Q=2/3Q=-4/3 pp  tt  101 ± 10295 ± 17 pp  tt ; t  Wb  6.2 ± 2.52.4 ± 1.5 Total background 38 ± 6 Determine charge of b-jet and combine with lepton  Use di-lepton sample  Investigate ‘wrong’ combination b-jet charge and lepton charge  Effective separation b and b-bar seems to be possible  Study systematics in progress pT(  ) events 10 fb-1

26. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 26 Top quark FCNC decay GIM suppressed in the SM Higher BR in some SM extensions (2-Higgs doublet, SUSY, exotic fermions) 3 channels studied in ttbar production: BR in SM2HDMMSSMR SUSYQS t  qZ ~10 -14 ~10 -7 ~10 -6 ~10 -5 ~10 -4 tqtq ~10 -14 ~10 -6 ~10 -9 t  qg ~10 -12 ~10 -4 ~10 -5 ~10 -4 ~10 -7 Preselection ≥ 1 lepton (pT > 25 GeV and |h| < 2.5) ≥ 2 jets (pT > 20 GeV and |h| < 2.5) Only 1 b-tagged jet ETmiss > 20 GeV

27. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 27 FCNC: Results E.g. t→qZ→ql + l - BR 5  sensitivity: Results statistically limited  Sensitivity at the level of SuSy and Quark singlet models L(fb-1) t  qZtqtq t  qg 105.1x10 -4 1.2x10 -4 4.6x10 -3 1001.6x10 -4 3.8x10 -5 1.4x10 -3  With 10 fb -1 already two orders of magnitude better than LEP/HERA/TEVATRON M jl+l-

28. Test the t  bW decay vertex Measure W polarization (F 0, F L, F R ) through lepton angular distribution in W cm system: Syst ( E b-jet,m top,FSR )  F 0  F 0 ~ 2% ;  F R ~ 0.01 Semilep. + Dileptonic cos(  l *) 0.000 0.297 0.703 SM  0.003  0.024FLFL  0.003  0.012FRFR  0.004  0.015F0F0 Error (±stat ±syst) (M t =175 GeV) ( hep-ex/0508061 ) L=10fb -1 Probing the Wtb vertex Angle between lepton in W rest frame and W in top rest frame

29. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 29 W polarization in top decay )  From W polarization, deduce sensitivity to tWb anomalous couplings  model independent approach, i.e. effective Lagrangian and 4 couplings (in SM LO ±1   2  limit (stat  syst) on = 0.04  3 times better than indirect limits (B-factories, LEP)  Less sensitive to and already severely constrained by B-factories F0F0 Anomalous coupling

30. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 30 tt spin correlation  Measurement of tt spin correlation (NP B690 (2004) 81) angle btwn spin analyzers direction in the t(t) rest frame angle between daughter and top spin axis s Wbl +,d,sv,u,clej *  (NLO) 0.40-0.401.-0.310.47 spin analyzing power * lej = least energetic jet in top rest frame In SM with M top  175 GeV,  (t)  1.4 GeV »  QCD Top decays before hadronization, i.e. ‘bare quark’ decay: substantial ttbar spin correlations predicted

31. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 31 Measurement method Rome + Pol4 fit Atlfast fit Correction function for A semileptonic Rome + Pol1 fit Atlfast fit Correction function for A D semileptonic  Selection cuts and reconstruction distort the parton level distribution (±stat) SM A 0.40  0.020.50  0.10 0.42 ADAD -0.28  0.01-0.29  0.06 -0.29  Results after selection, reconstruction and correction: Fast SimFull Sim 9*cos  1 *cos  2 1/weight 3*cos  1/weight  Good agreement Full Sim. / Atlfast  Syst. dominated by b-JES, top mass and FSR  ~4% precision on spin correlation parameters  Tevatron expectations (2 fb -1 ):  A stat /A~40%

32. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 32 Search for resonances Many theoretical models include the existence of resonances decaying to top-topbar  SM Higgs (but BR smaller with respect to the WW and ZZ decays)  MSSM Higgs (H/A, if m H,m A >2m t, BR(H/A→tt)≈1 for tanβ≈1)  Technicolor Models, strong ElectroWeak Symmetry Breaking, Topcolor, “colorons” production, […] Study of a resonance Χ once known σ Χ, Γ Χ and BR(Χ→tt)  Reconstruction efficiency for semileptonic channel:  20% m tt =400 GeV  15% m tt =2 TeV 1.6 TeV resonance M tt  xBR required for a discovery m tt [GeV/c 2 ] σxBR [fb] 30 fb -1 300 fb -1 1 TeV 830 fb

33. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 33 Implicit new physics.. E.g. SUSY particles modify ttg vertex, investigate M tt :  All SuSY masses  M SuSY  M  400 GeV, “SuSY Yukawa” might be visible in higher order corrections Diagrams for SuSy corrections to gg  ttbar Modified Mtt distribution for various tanb - Can be measured with ~20% precision using 10fb -1 data - QCD NLO effects estimated with MC@NLO Confirmation of SUSY?

34. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 34 Single top production Direct determination of the tWb vertex (=V tb ) Discriminants: - Jet multiplicity (higher for Wt) - More than one b-jet (increase W* signal over W- gluon fusion) - 2-jets mass distribution (m jj ~ m W for the Wt signal only) Three production mechanisms: It is likely that single top production will be seen for the first time at LHC t-channel 245±27 pb Wt-mode 62.2 pb s-channel 10.2±0.7 pb Powerful probe of V tb ( dV tb /V tb ~1% @ LHC ) An effective way to study new phenomena FCNC (t-ch.), new gauge bosons or KK excitations (s-ch.), H ± tb … Background to tt, WHlnbb, some SUSY and BSM final states Probe various regions on Wtb coupling, with q 2 0, q 2 =0 Constraints on the b-parton density Source of polarized top-quarks See talk of Akira Shibata See the talk of Teh Lee Cheng

35. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 35 NLO process in MC@NLO NLO processes on s-and t-channel included (Frixione, Laenen, Motylinski)  Wt-mode and spin effects being worked on  Separation in s- and t-channel becomes matter of convention  MC@NLO is improving both NLO calculations and standard MC simulations Validation in Atlas ongoing P T of the B-hadron

36. Common feature: 1 lepton, pT>25GeV/c High Missing E T  2 jets (at least 1 b-jet) t-channel: exactly 2 jets (1 b-jet), window cut on m lnb Stat: 7000 events (S/B=3) Syst: dominated by E b-jet and Lum. Error Back: tt, Wbb and W+jets (  <1.5%) Wt-channel: exactly 3 jets (1 b-jet), window cut on m lnb Stat: e~1% (S/B=15%) (  ~ 4%) (  ~7-8%) L=30fb -1 Production cross section s-channel: exactly 2 jets (1 b-jet), window cut on m lnb Stat: 1200 events for tb (S/B=10%) Syst: E b-jet, Lum. Error, back X-section ttbar production and t-channel are background

37. Stan Bentvelsen - ATLAS-UK - Sept 19, 2006 P 37 Conclusions Top quarks for commissioning the detectors  Top peak should be visible with eyes closed Precise determination M top is waiting…  Challenge to get  M top ~ 1 GeV Is the top-quark a SM particle?  Measure V tb, charge, CP, spin, decays  Single top production Today’s signal, tomorrow’s background  Top quarks as main background for many new physics channels  New physics decaying into, or associated with tt-bar LHC is top factory  (tt)~830 ±100 pb -1 10 7 events in first (?) year