Top Quark Physics at Hadron Colliders

Transcription

1 Top Quark Physics at Hadron Colliders Arnulf Quadt LHC at CERN Tevatron at Fermilab Seminar at 26. April 2007 TU TUDresden DresdenSeminar PaPag ge e1 1

2 Outline Introduction... Theory & Experiments... Top Quark Production... in Weak and Strong Interactions... Top Quark Properties Mass & Charge Decay (Lifetime, W helicity,...) Conclusion Page 2

3 Introduction Page 3

4 Top Quark in the Standardmodel Why is the Top Quark so interesting? completes the quark sector mtop ~ 180 GeV / c2 large mass short lifetime sensitive to physics beyond the Standard Model Only `recent' discovery charge, mass,...? weak and strong top physics τ ~ s Λ 1QCD Tevatron LHC Higgs Boson coupling to fermions : mt ~ v/ 2, Yukawa coupling λ t~1 g ~ mf Page 4

5 Top Quark Physics Tevatron Run I : top quark discovery (1995) Run II &LHC : with high precision answer... Why is top so heavy? Is top/third generation special? Is top involved in EWSB? Is it connected to new physics? Precision measurement of couplings tt Production Cross Section tt Production via interm. Resonances EW single top production Production Kinematics Spin Polarization Top Mass Top Width Top Spin Top Charge t Anomalous Couplings CP Violation Rare/non SM Decays Branching Ratios Vtb b i g t 1 5 V tb b W 2 2 g 0.67 V tb 1 mtop mw phase space W helicity W l ν Page 5

6 The TEVATRON at Fermilab Chicago CDF DØ Equivalent lumi for CDF km Bo os ter Teva tron _ p source ~ 2.38 fb 1 delivered p _ p ~ 2.00 fb 1 recorded Ma in Inj ect or & Recycl er _ p s = Te V, t = 39 6 ns p Run I 1987 (92) 95 Lint ~ 125 pb 1 Run II (?) 4 9 fb 1 CDF & DØ data taking ε ~90% Page 6

7 The TEVATRON Performance World (ISR) record cm 2 sec 1 Run I record Peak luminosity: _ P, twice the SPPS number! in close to or above optimistic design scenario recycler and electron cooling in operation Page 7

8 The CDF and DØ Experiment new bigger silicon, new silicon and fibre tracker new drift chamber, TOF new ~2 T solenoid Upgraded calorimeter and upgraded muon system muon system upgraded (track) trigger/daq Upgraded DAQ/trigger Roman pots Displaced track trigger resolutions: 19 countries, 83 institutes, ~750 physicists EM: σ E/E = % / sqrt(e) 664 physicists HAD: σ E/E = % / sqrt(e) Page 8

9 The Large Hadron Collider LHC CMS ATLAS The Large Hadron Collider: _ proton proton collider (no p) 2 separate beampipes first collisions in 2007 high energy: s = 14 TeV 40 Mio. collisions per second 4 experiments: ATLAS, CMS, ALICE, LHC B 10 fb 1 per year LHC dipoles LHC quadrupoles Page 9

10 Top Quark Production in Strong Interactions Page 10

11 Decay Topology in Top quarks decay predominantly (~100%) to a W Boson and a b quark τ +X µ+jets e+jets e+e e+µ µ+µ Top Antitop Signatures: `dilepton channel' 5% : 2 jets, 2 charged leptons, 2 ν t t 44 % hadronic top decay 22 % 14.8 % 1.4 % 1.4 % 14.8 % 2.8 % 'lepton+jets channel' 30%: 4 jets, 1 charged lepton, 1 ν 'all jets channel' 40%: 6 jets always 2 jets are b jets Page 11

12 q t _ q _ t g _ t g _ t g _ t g t g t g t Tevatron LHC _ qq ~ 85 % 15% gg ~ 15 % Run I Run II 85% LHC (2 fb 1) (10 fb 1) nr ttbar interaction rate Strong Top Quark Production 8 * 10 6 σtt (mt sample 1 b tag) establish top signal measure cross section as QCD test cross section and topology close to Higgs physics LHC Tevatron Page 12

13 Ttbar Xsec Measurements at Tevatron dilepton l+jets alljets Topological selection (lepton pt, MET, Njets) counting experiment b tag selection (lepton pt, MET, SVX tag, Njets) counting experiment lepton+track (lepton pt, MET, isolated track, Njets) counting experiment dilepton 2 dim. (MET,Njet) fit for ttbar, WW, Z τ τ 2 dim. fit Topological selection (e/μ +jets, lepton pt, MET, Njets), topological & kinematic variables, 1 dim. fit 1 dim. fit b tag selection (e/mu+jets, lepton pt, MET, Njets), b tag (SVX, IP, jet prob., soft μ ) counting exp. kinematic fit (MET or jet ET) in b tagged events 1 dim. fit combined fit of 0, 1, 2 tag sample and B(t Wb)/B(t Wq) 2 dim. fit Kinematic & topological selection, Njet distribution 1 dim. fit Kinematic & topological selection, ANN output counting exp. Page 13

14 Run II Top Cross Section Summary status Oct errors between different channels are correlated status Oct Measurements demonstrate success of various top detection techniques Results within errors consistent with NNLO SM prediction for 1.96 TeV of ~7 pb 1 Combination being worked on (TevEWWG) Latest results (760 pb 1) achieve ~15% precision Forward backward charge asymmetry (8%) first thesis result Page 14

15 Top Quark Production in Weak Interactions Page 15

16 Single Top Quark Production V tb V tb Page 16

17 Motivation of Single Top Quarks Directly measure Vtb for the first time Cross section sensitivity to processes beyond the SM (W',...) Source of polarized top quark spin correlations Important background to Higgs search Demonstrate techniques to extract small signal out of large background Similar topology to tt in l+jets, BUT fewer jets & more forward background (W+jets, tt, dibosons,...) Page 17

18 A Real Challenge... W tt single top 2 :=1 Page 18

19 Single Top Quark Production no b tagging yet W +jets fakes (m ultijet) W +jets fakes (m ultijet) single top single top M issing E T [G ev ] pt jet1 [G ev ] signal < background uncertainty! Page 19

20 Multivariate Analysis Techniques Likelihood discriminants (CDF) Artificial neural network (CDF) Matrix element (DØ, CDF) Bayesian neural network (DØ) Boosted decision trees (DØ) background signal Page 20

21 CDF Results 1 fb 1 Likelihood discriminants: σ < % CL 2 artificial neural networks: σ < % CL matrix element: σ = pb (2.3σ ) 1.3 but: use same data selection 0.7% compatibility of all analyses Page 21

22 Boosted Decision Tree Output 49 variables single top W+jets fakes (multijet) top pair Page 22

23 Boosted Decision Tree Output 49 variables single top W+jets fakes (multijet) top pair HT < 175 GeV Page 23

24 Boosted Decision Tree Output 49 variables single top W+jets fakes (multijet) HT < 175 GeV top pair HT > 300 GeV Page 24

25 Boosted Decision Tree s+t Observed Result σ = 4.9 ± 1.4 pb background com patibility:0.035% (3.4σ) SM com patibility:11% (1.1σ) Page 25

26 Evidence for Single Top 1 fb 1 3.4σ excess prob. 3.5E 04 December 2006 Page 26

27 First Direct Measurement of Vtb 1 fb 1 V tb previously only indirectlim its: V tb = ± (1σ C L ) CKM Fitter Group for Beauty 2006 assume: Vtd 2 + Vts 2 «Vtb 2 assume: pure V A and CP conserving Wtb interaction no assumption on quark families or CKM matrix unitarity 0.68 < V tb 1 (95% C L) Page 27

28 Summary: Evidence for Single Top Production hep-ex/ fb 1 first evidence for single top quark production (DØ decision trees) σ = 4.9 ± 1.4 pb (3.4σ) first direct measurement of Vtb (DØ decision trees) 0.68 < V tb 1 (95% C L) CDF analyses have good sensitivity but got unlucky... Page 28

29 Top Quark Properties... Charge & Mass... Page 29

30 Top Quark Charge (I) SM: Top Charge +2/3 e Other models predict 4/3 e Standard Model : + 2/3 e b 1/3 e Strategy: measure σ (t+γ LHC (10 fb 1) b jet Tevatron (here) Alternative Model : e.g. exotic 4th generation + Higgs triplet S.Chang et al., Phys Rev D59, (1999) 4/3 e t b 1/3 e Q +1e +1e W 0e 1e ν 1e l Analysis : a) associate lepton and b quark to top quark use a kinematic fit for ttbar hypothesis b) determine charge of b jet pt weighted sum of charged tracks W 0e ν l Present Z ll and Z bb data not inconsistent with 4/3 e top quark of mass 270 GeV/c2 associated to a b jet Page 30

31 Top Quark Charge (II) First measurement of this fundamental top quark quantum number : DØ Run II (365 pb 1) 17 candidat events with two tagged b jets, lepton, missing ET, 4 jets two entries per event for top and anti top discriminate b and bbar with jet charge algorithm i q i p q jet = i P 0.6 Ti 0.6 Ti corrected for B mixing and, pt > 0.5 GeV & Δ R > 0.5 b sample charm fraction calibrate Monte Carlo with data using two jet heavy flavor sample with opposite jet tagged with μ flavor exclude the hypothesis of an exotic quark with charge = 4/3 e at 94% confidence level Page 31

32 Top Quark Mass Measurements at Tevatron dilepton l+jets Neutrino weighting (η ϕ ) 1 dim. fit Phi weighting (ϕ η ) 1 dim. fit Pz(tt) method 1 dim. fit ME weighting 1 dim. fit ME method 1 dim. fit Template method in mtop after kinematic fit, topological or b tag, with internal or external JES constraint 1 or 2 dim. fit Matrix Element/Dynamical Likelihood Method, topological or b tag, with internal or external JES constraint, complex analysis 1 or 2 dim. fit Ideogram method (W LEP), compare signal and background mass spectrum, chi^2 weighting (kine fit), with internal/external JES constraint 1 or 2 dim. fit Decay Length Method, compare transv. Decay length spectrum with expecation from cτ (B) β (mtop)γ (mtop) alljets Kinematic fit, only from Run I, little sensitivity 1 dim. fit 1 dim. fit Page 32

33 CDF Template Analysis in L+Jets Data χ 2 mass fitter: Finds top mass that fits event best One number per event Additional selection cut on resulting χ 2 Dataset 680 pb 1 Wbb MC tt MC Mass fitter Signal/background templates Likelihood fit Data (mtop, Δ JES, nb, ns) Result Likelihood fit: Best signal + bkgd templates to fit data with constraint on background normalization Page 33

34 CDF Template Analysis in L+Jets Fit four data samples (0 tag, 1 tag(loose), t tag(tight), 2 tag with SecVtx tagger) in mtop and Δ JES, i.e. 2 dim fit : mtop =173.4±1.7 stat ±1.8 JES ±1.3 syst. GeV / c 2 mtop =173.4±2.8 GeV / c 2 Page 34

35 Dilepton Channel: Neutrino Weighting 2 neutrinos: kinematics underconstrained py ν ETmiss b q ν t W q ν no ν pz inform ation + b t px ℓ+ ℓ- W ν Page 35

36 Dilepton Channel: Neutrino Weighting 2 neutrinos: kinematics underconstrained py ν assume top mass ETmiss b t W q ν ν no ν pz inform ation + b t W ℓ+ take (anti ) neutrino η from MC ℓ- entries q px ν neutrino η Page 36

37 Dilepton Channel: Neutrino Weighting 2 neutrinos: kinematics underconstrained py ν assume top mass ETmiss b t W q ν ν no ν pz inform ation + b t W ℓ+ take (anti ) neutrino η from MC ℓ- entries q px ν sum over all solutions neutrino η Page 37

38 Neutrino Weighting Algorithm compare measured MET with expected (MC) for different mtop hypotheses derive w(m ) for every event w eight top m top (G ev ) Page 38

39 Neutrino Weighting Algorithm compare measured MET with expected (MC) for different mtop hypotheses derive w(m ) for every event w eight top rm s extract tem plates m top (G ev ) m ean Page 39

40 Neutrino Weighting Algorithm: Result NEW: simultaneous 3 (2) dimensional fit to signal (background) templates rms=45 GeV 1 fb 1 57 events m top m ean maximum Likelihood function L(mtop,mean,rms): PRELIMINARY m top=172.5 ± 5.8(stat.)± 5.5 (syst.) G ev Page 40

41 Matrix Element Method Obtain event probabilities by folding differential X section with object resolutions: Proton W g W Anti Proton b Jet l ν Lepton q' _ q _ b P m x, m t = Acc x missing ET Jet Jet Jet select exactly 4 jets sum over jet parton combinations energy resolution for μ, light quark and b jets 1 n dq dq f q f q d σ y ; m t W x, y σ Acceptance (selection, trigger,...) & normalization PDF s LO Matrix element detector resolution phase space & Δ JES = diff. Xsec Pbkg(x) accordingly... event probability Pi Multiply Pi's for all events and maximize wrt mtop Page 41

42 Matrix Element Method probability densities for every event as function of mtop Pm signal signal Pm mtop q b t + t W b background Pm mtop W q Pm signal mtop ν experim ent Pm = mtop mtop W mass constrains jet energy scale ℓ +q q' Page 42

43 DØ Top Mass in L+Jets maximum Likelihood fit using signal and background pdfs 1 fb events at least one jet has b tag m top m top = ± 1.6 (stat)± 1.4 (JES)± 1.4 (syst) G ev Page 43

44 Top quarks at Tevatron produced nearly at rest boost of the b quark a function of mtop mt mb m W mt b = mt mb mb signal sample Measure transverse decay length of B hadrons from top decay infer on top quark mass control sample Decay Length Technique Select l + 3 jets events with 1 SecVtx tag in 695 pb pos. SecVtx tags in 375 events mtop = stat. ±5.6 syst. GeV /c 2 m top JES =0.3 GeV / c Page 44

45 Top Mass Summary/Combination systematics limited: jet energy scale ttbar modeling W+jets modeling 1 fb 1 TevEWWG (Mar/2007): status Oct hep ex/ Page 45

46 Top Mass Outlook only D0, in l+jets Tevatron could reach Δ mtop=1 GeV/c2 in combination of all channels and both experiments if... ultimate precision in top mass from LHC and Tevatron expected to be comparable Page 46

47 Top Quark Physics Outlook CDF + DØ combined expected precision with 2 fb 1 ATLAS/CMS separate with fb 1 (mass precision from total lumi) CDF+DØ W helicity F0, F+ ATLAS/CMS CDF+DØ 0.09, R2b/1b 4.50% (stat.error) ~0.2% Vtb from R... B(t γ q) 2 * *10 4 Δ mtop 1 2 GeV 1 2 GeV Δ mw 25 MeV 20 MeV ATLAS/CMS single top 20.00% 0.71% Γ t from single top 25.00%... Vtb from single top 12.00% 0.36% *10 4 discovery? discovery! 12 16% B(t Zq) Higgs Yukawa Coupl. yt LHC dominates in statistics limited analyses Tevatron will play a role in systematics limited measurements Select events with lepton (pt>20 GeV) exactly 4 jets (ET > 40 GeV) mass of 3 jets with highest vector pt in 150 pb 1 (a few days of LHC) gives: use top physics for LHC commissioning m Page 47

48 Summary The top quark is unique among the six quarks mass charge spin... production (strong & weak) decay couplings (gauge or new)... Do/will study the top quark and its role in Nature with unprecedented precision... Tevatron and Experiments performing well Top physics sensitivity as expected or better (mass) Searches statistics limits (better at LHC) measurements already now systematics limited Page 48

49 Backup Page 49

50 Further Top Quark Properties Page 50

51 Weak Top Decay Helicity of the W V A i g t 1 5 V tb b W 2 2 Top Standard Model weak decay V A coupling as it is for all the other fermions W t W0 fraction F W+ spin=1 W t W W Left Handed W- t spin =1/2 b sum b spin =1/2 t b W0 Longitudinal fraction F0 Suppressed by the V-A coupling b W+ Right Handed fraction F w cos l b = F 1 cos l b F 0 1 cos l b F 1 cos l b In SM Helicity of W manifests itself in decay product kinematics 2 F = m 2W M 2top 1 2 m 2W F 0= 1 2 M 2top m2w M 2top 2 F m b / mt 0... for any V, A combinations... Page 51

52 W helicity in Top Quark Decays Charged leptons from left handed W opposite to W topological + b tag sample in l+jets kinematic fit + decay angle reconstr. soft pt (helicity angle in W rest frame)... longitudinal W transverse to W hard pt l+jets with 1, 2 b tag and di lepton F0 = present status: F0 = F+ < C.L. F+ < 95% CL 230 pb 1... no deviations from SM predictions eventually simultaneous fit for F0 and F+... Page 52

53 Ttbar Production via Resonances Does something new (narrow resonance) produce ttbar pairs? Lepton + 4 jets with 1 b tag in 370 pb 1 kinematic fit to ttbar hypothesis assume SM rate for SM ttbar production no significant excess observed Lepton + 4 jets (no b tagging) in 680 pb 1 use matrix element technique assume SM rate for SM ttbar, diboson, QCD assume everything else is W+jets no significant excess observed Interpretation of σ X BR(X tt) limit CDF: in terms of mass limit of a Z' in topcolor assisted technicolor (hep ph/ ) mx > 725 GeV DØ: mx > 680 C.L. Page 53

54 Search for Charged Higgs In 2 Higgs Doublet Models (e.g. MSSM) H± exist If mh± < mt t H±b can compete with t Wb Changes in event topology according to H± decay (scalar, channels) : large tan β small tan β : ± excess of τ decays in ttbar events ± excess of fully hadronic ttbar events 2 extra b jets in ttbar events H { H c s ± H W h W b b Combine CDF data (192 pb 1) in dilepton, lepton+jets (1 b tag), lepton+jets ( 2 b tags), lepton+hadronically decaying τ taking t H±b and H± decay branching ratios into account Page 54

55 Search for Charged Higgs tan β >30: Br(t H±b) enhanced by low μ tan β >30: Br(t H±b) suppressed by large μ tan β ~ 1: H± Wh0 occurs If Br(H± τν ) = 100% limit on Br(t H±b)<0.4 Scan over all Br(H±.. ) quote weakest limit on Br(t H±b)<0.91 (model independent) Page 55

56 Top Quark Lifetime First direct limit on the lifetime of the top quark top quark lifetime in SM s (cτ = μ m) search for anomalous top production via new, long lived particle lepton + 3 jets with 1 b tag in 318 pb 1 97 e+jets candidates 60 μ +jets candidates measure impact parameter d0 for lepton tracks use max. likelihood fit with templates of arbritrary lifetime (incl. track resolution) Best fit: cτ = 0 μ m upper limit: cτ < 53 μ m (@ 95% CL) Page 56

57 Available Results & Ongoing Studies Tab le of T op Pr ope rty Re sults : Measurement Mass Cross Section W Helicity W Helicity Top Charge Resonance Searches Top Lifetime 4th Generation t' Quark Charged Higgs Searches Anomalous Kinematics BR(t->Wb) / BR (t->wq) Single Top Single Top Spin Correlation Best Results Dataset /- 2.8 GeV/c /- 1.1 pb F0 = F+ < 95%CL > 95%CL s-channel: σ (tb) < 5.0 pb t-channel: σ (tqb) < 4.4 pb κ > pb pb pb pb pb pb pb pb 194pb-1 194pb-1 162pb-1 370pb-1 370pb-1 125pb rule out +4/3 94%CL Mx0 < 725 GeV/c2 cτ < 53 µ 95%CL 196 < M(t') < %CL Limits on BR(t->H+b) no high pt excess... uction a symme try Strong prod -1 Page 57

58 Topological Xsec in L+Jets (I) input examples Select l+ 4 jet events in 230 pb 1 choose topological variables: with strong separation potential with small sensitivity to jet energy scale use the following 6 variables: angular dependent: sphericity aplanarity centrality energy dependent quantities: HT Kt,min Background sensitive quantities: Δ φ (l,met) topological likelihood: i S i P= i S i i Bi i= 1..6, S = tt-distr ib ut ion, B = Wjj jj-d is tri bu tion Page 58

59 Topological Xsec in L+Jets e+jets μ +jets Combined result: tt = stat. 1.1 syst. ±0.4 lumi pb Systematic uncertainties dominated by jet energy scale uncertainty shape uncertainty, W+Jets Q2 scale Similar CDF ANN analysis with 760 pb 1 tt =6.0±0.6 stat. ±0.9 syst. pb Page 59

60 B Tag Xsec in L+Jets Detector and physics simulation of b decay not sufficiently precise measure b tagging efficiency and mistag rates in data and weight MC accordingly Select (W l ν ) + jet events in 695 pb 1 Count events in jet multiplicity bin (counting experiment) SecVtx tagger Single Tag: t t P tag =53.4 % W light P tag =1.7 % Double Tag: t t P Dtag =16 % tag untagged N expected = N t t t t t t P tag bgd untagged N bgd bgd P tag Page 60

61 B Tag Xsec in L+Jets 1 b tag and HT > b tag GeV 1.2 tt =8.2±0.6 stat. ±1.0 syst. pb Systematics dominated by b tagging tt = stat. 1.3 syst. pb Similar analysis by DØ (365 pb ): =8.1 stat. syst. ±0.5 lumi tt 1.2 pb Page 61