1 / 36

Measurements of Single-Top Production at the LHC

Measurements of Single-Top Production at the LHC. Wolfgang Wagner Bergische Universität Wuppertal on behalf of the ATLAS and CMS collaborations. Top Physics Workshop Sant Feliu de Guixols , Spain September 28, 2011. Content: Introduction t -channel measurements (ATLAS and CMS)

taipa
Télécharger la présentation

Measurements of Single-Top Production at the LHC

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Measurements of Single-Top Production at the LHC Wolfgang Wagner BergischeUniversität Wuppertalon behalf of the ATLAS and CMS collaborations Top Physics Workshop SantFeliu de Guixols, Spain September 28, 2011 Content: Introduction t-channel measurements (ATLAS and CMS) Wt searches (ATLAS and CMS) Status of s-channel search (ATLAS) Search for FCNC single-top production (ATLAS) Summary

  2. 1) Introduction top-quark production via the weak interaction. s-channel t-channel associated Wt production cross sections at LHC with √s = 7 TeV (mt = 173 GeV) 64.2 ± 2.6 pb 15.6 ± 1.3 pb 4.6 ± 0.2 pb cross sections at the Tevatron with √s = 1.96 TeV (mt = 173 GeV) 2.1 ± 0.1 pb 0.25 ± 0.03 pb 1.05 ± 0.05 pb Calculations by N. Kidonakis: arXiv 1103.2792, 1005.4451, 1001.5034at NLO + NNLL resummation (NNLOapprox)

  3. Test of the SM prediction. Does it exist? ✔ Establish different channels separately. Cross section  |Vtb|2Test unitarity of the CKM matrix, .e.g.Hints for existence of a 4th generation ? Test of b-quark PDF: DGLAP evolution Search for non-SM phenomena Search W’ or H+(Wtor s-chan.signature) Search for FCNC, e.g. ug t … Why look for single top-quarks? • Single top as an experimental benchmark • Object identification: lepton fake rates, QCD background estimates, b-quark jet identification, … • Redo measurements of top properties in different environment, for example, Mtop, W polarization in top decay, …

  4. Overview of performed analyses t channel Wt channel s channel lepton+jets and dilepton channel: cut-based cut-based • cut-based • neural network

  5. 2) t-channel analyses: documentation • Largest cross section of single-top processes • Improved S/B ratio (10%) compared to Tevatron (7%) Analysis history at ATLAS • CONF note (Moriond) with 35 pb-1 (1.6 s),ATLAS-CONF-2011-027 • CONF note (PLHC) with 156 pb-1 (6.2 s),ATLAS-CONF-2011-088 • CONF note (EPS) with 0.70 fb-1 (7.6 s), ATLAS-CONF-2011-101 ATLAS-CONF-2011-101

  6. t channel event selection: leptonic W decay • Data sets defined by single lepton (e / m) trigger • Charged lepton selection (electron / muon): • pT> 25 GeV, pT (m) > 20 GeV, ET(e) > 30 GeV • |h(m, e)| < 2.5 • Relative isolation • Missing transverse energy • ETmiss> 25 GeV • QCD multijet veto • MT(W) > 60 GeV – Etmiss • Select only events with leptonic W decays, to suppress QCD-multijets background. • Some acceptance due to W ® tndecays.

  7. t channel event selection: jets • Jet definition • Anti-kT algorithm R=0.4, R=0.5 (particle flow) • pT > 25 GeV • |h| < 4.5, Measurement of forward jets is crucial to t-channel analyses. • b quark jet identification • Exactly one secondary vertex tag(in 95% of all 2 jets events the b quark jet from top decay is tagged) • Number of jets • NN analysis : 2 • cut based: 2 & 3

  8. Background processes (before b-tagging) single top t-channel (~1%) top-antitoppair production (~1%) mainbackground: W+jetswithseveralcomponents: W + light jets (55 – 65%)W + charmjets (~20%)W + bottomjets (2 – 3%) QCD multijets(fake lepton) background (5 – 10%)

  9. Background estimation - strategy • Top-antitop pairs, Wt, s-channel, diboson, Z+jets • MC normalized to theoretical (or measured) cross-section • Acceptance / efficiency obtained by Monte Carlo Monte Carlo based backgrounds • instrumental backgroundreliableestimationonlyfromdata • reduceasmuchaspossibleQCD veto • fit discriminant; ATLAS: Etmiss • datadriveneventmodelformultivariate methods:jet-electronmodel QCDmultijetsbackground • Alpgen (ATLAS) LO+LL prediction • data driven scale factors W+jets

  10. Multijet background estimate ATLAS: fit Etmiss cut value • jet-electron model works also well for muons • Fraction of multijets background: 5 – 10% • Systematic uncertainty: ± 50%

  11. W + jets background estimate ATLAS Cutbased analysisCalculate scale factors kcc/bb, kWc, klightbased on event yield in 1-jet and 2-jet tagged sideband and 2-jet pretagged data set NN analysisFit NN output for single-top and backgrounds simultaneously Overall ALPGEN andMadgraphmodelsworkquitewellwithinuncertainties.

  12. Systematic uncertainties: background rates • Theoretical cross-section uncertainties *includes Berends scaling and HF uncertainty • W + jets and multijets normalization to data

  13. Uncertaintieson object and kinematic modeling • Detector simulation and object modeling • Jet energy scale, jet energy resolution Leptons: trigger, identification efficiencies, energy scale , lepton energy resolution • B-tagging / mistagscale factor uncertainty • Monte Carlo generators • ISR / FSR • t-channel (ATLAS): MCFM vs. AcerMC • ttbar(ATLAS): MC@NLO vs. Powheg • PDF: CTEQ6.6 vsMSTW08 • Q2 scale for W+jets • Pile-up modeling • Luminosity • ATLAS: 4.5%

  14. T-channel Cut-based analysis at ATLAS • Cuts are optimized including systematicsstrong reduction of jet energy scale uncertainty • Counting experiment • Uses 2 and 3-jet channels • Separation in channels lepton charge and flavor optimize statistical power • Statistical method: profile likelihood fit measured cross section: s(t-ch) = 90 ± 9 (stat.) +31-20 (syst.) pb Observed significance 7.6 s(expected: 5.4s) SM: st = 64.2 ± 2.6 pb • Dominating syst. uncertainties: • B-tagging: +18 / -13% • ISR / FSR: ± 14%

  15. Neural network analysis Idea: Combine many variables including correlations in one discriminate 13 input variables 33 nodes in hidden layer

  16. t-channel neural network analysis at ATLAS • Signal already well visible in Mlnb • 13 input variables • Training: 50% signal, 50% background. • Maximum likelihood fit to NN output distribution.simultaneous determination of background rates • Frequentist method to estimate systematic uncertainties. Observed cross section: s (t-ch.) = 105 ± 7 (stat.) +36-30 (syst.) pb SM: st = 64.2 ± 2.6 pb • Dominating syst. uncertainties: • Jet energy scale: +32 / -20% • B-tagging: ±13% • ISR / FSR: ± 13%

  17. 3) Wt analyses Two channels according to W decay modes: Dilepton channelboth W:W  enor W  mn2 charged leptons, ETmiss, 1 b-jet Lepton + jets channelW  enor W  mn+ W  qqbar 1 chargedlepton, ETmiss, 3 jets • CONF note with 35 pb-1(Moriond)ATLAS-CONF-2011-027 • CONF note with 0.70 fb-1 (EPS)ATLAS-CONF-2011-104

  18. Event selection in the dilepton channel • Lepton selection (electron / muon): • pT > 25 GeV (ATLAS), |h|< 2.5 • Relative Isolation • Exactly two leptons (ee /mm/ em) • Jets • pT> 30 GeV • ATLAS: |h| < 2.5, • Exactly one jet. : No b- tagging! • Missing transverse energy • ETmiss> 50 GeV (ATLAS) • Z-mass veto (ee/mm –channel) • |M(ll)-M(Z)| > 10 GeV, CMS: M(ll) > 20 GeV • Z → ttveto (ATLAS) • DF(l1, Etmiss) +DF(l1, Etmiss) > 2.5 • Kinematic selection at CMSem channel: HT > 160 GeV,pT(lljn=system) < 60 GeV

  19. Dilepton backgrounds – estimation strategy Monte Carlo based dibosonWW / WZ / ZZ Drell-Yan Z/g ll ABCD method W + jets(lepton fakes) matrix method (ATLAS) top-antitopdilepton channel normalization in W + 2 or more jets side bandCMS: exclusive W + 2 jets sample

  20. Top-antitop background and kinematic modeling signal region top-antitop sideband region simultaneous fit (similar technique in ATLAS) Good agreement with expected jet multiplicity distribution and kinematic distributions.

  21. Wtdilepton analyses’ results Event yield after final selection (Njet = 1): Final event yield for 2.1 fb-1 at CMS: • Observed significance: 2.7s(1.8s expected) • Observed cross section:sWt= 22 +9-7 (stat.+ syst.) pbSM: sWt= 15.6 ± 1.3 pb Observed cross section (significance 1.2s): sWt= 14.4 +5.3-5.1 (stat.) +9.7-9.4pb Observed limit @ 95% C.L. sWt< 39.1 pb Significances are determined with maximum likelihood ratio:

  22. Wt lepton + jets channel • Experimental signature: • Isolated charged lepton • Missing transverse energy • Three high-pT jets Event selection very similar to t-channel analysis, same background estimation strategy Analysis of 2010 data with 35 pb-1 • ATLAS-CONF-2011-27 (Moriond 2011) • Obtain S/B = 4 – 6% • Dilepton and lepton+jets channel were combined:observed limit at the 95% C.L.:s (Wt) < 158 pb • Multivariate analyses are in preparation.

  23. 4) Search for s-channel production • Smallest cross section of all single-top processes.(antiquarks in the initial state needed) • Signature similar to t-channel, but: • No forward jet. • Two central b-quark jets. • Jet definition uses: |h| < 2.5. • Use double tagged events. • First s-channel analysis at ATLAS using 0.70 fb-1. ATLAS-CONF-2011-118 Cut-based analysis

  24. Limit on s-channel production Statistical analysis:Profile likelihood Event yield after final selection: Observed limit @ the 95% C.L.: ss-channel < 26.5 pb SM: ss= 4.6 pb

  25. Summary • Single top t-channel production has been observed at ATLAS (7.6s @ 0.7 fb-1 • Measured t-channel cross sections are in agreement with the SM (64.2 ± 2.6 pb).s(t-ch.) = 90 ± 9 (stat.) +31-20 (syst.) pb • With 0.70 fb-1 (ATLAS) already systematically (~30%) limited (stat. unc. 10%). • First steps to measure subleading single-top processes: • s(Wt) < 39 pb @ 95% C.L. • s (s-chan.) < 26.5 pb @ 95% C.L.

  26. Thank You!

  27. Backup

  28. 5) FCNC in top-quark production • FCNC: Flavor-Changing Neutral Currents • significant in extensions of SM (e.g. SUSY) • any evidence reveals new physics SM GIM mechanism BR → 10-13 Very effective in the top sector! hep-ph/0409342 At a hadron collider more effective to look for FCNC production than decay.

  29. FCNC single top quark signature FCNC decays can be neglected since very large couplings are already excluded: Best current limits by DØ:Phys. Lett. B 693 (2010) 81 SM dominant FCNC single top-quark production can be studied assuming SM decay of the top quark. • Same event selection as t-channel analyses, but • Use only central jets: |h| < 2.5 • Exactly one jet Protos generator used for signal modeling.

  30. Expected event yield for Lint = 35 pb-1 • Uncertainties include statistical and cross-section uncertainties. • Assumed signal cross section: 1 pb • Scale factors for W + jets processes are determined in a simultaneous fit to the NN discriminant (not included in the table above).

  31. b-jet identification b hadron lifetime:   1.5 ps c  450 m  typical decay length: O(mm)  lifetime based b-taggers impact parameter based tagger secondary vertex based tagger documentation on b-tagging in ATLAS: ATLAS-CONF-2011-102

  32. Monte Carlo samples All samples with full detector simulation using GEANT4. Pileup is simulated with 50 ns bunch trains.

  33. Modeling of Single-Top Events: Example 2nd b Quark from b-quark PDF flavourconservation (in strong interaction):2nd b fromshower MC (DGLAP evolution) Problem in MC@NLO + Fortran Herwig: Solution:matching of bu® td and gu®tdbbar processes • Switch of signal generator in ATLASfrom 2010 to 2011 analyseschange in acceptance: -20% • issue fixed in Herwig++

  34. W+jetsmodeling: ALPGEN andMadgraph • ALPGEN (M. Manganoet al.) @ ATLAS, Madgraph (Maltoniet al.) @ CMS • models multi-gluonemissionby LO matrixelements + partonshower LO+LL accuracy • workwith Pythia and Herwig • overlapbetween ME and PS must beremoved  MLM matching • „hard“jetsaremodeledby ME, „softer“jetsbytheshower MC • eachprocessismodeledbymanyspecific „parton“samples, forexampleW+bbbarWbb+0p with W e, Wbb+1p with W e, … • Overlapbetween heavy-flavorsamples must beremovedbyhand. Overall ALPGEN andMadgraphmodelsworkquitewellwithinuncertainties.

  35. Z → t+t-veto • Z → ttveto (ATLAS) • DF(l1,Etmiss) +DF(l2, Etmiss) > 2.5

  36. Wt systematic uncertainties Profile maximum likelihood fit: Statistical uncertainty: +37 / -35% Dominating syst. uncertainties:

More Related