1. QCD at the Tevatron Marek Zieliński University of Rochester Physics at LHC, Vienna, 16 July 2004

2. Physics at LHC, 16 July 2004Marek Zieliński, Rochester2 Outline l Introduction l From hard to soft QCD studies: èCone and kT jets èDijets èW + jets production èW/Z + b èDiphotons èDijet decorrelations èJet shapes èUnderlying event studies lSummary Not covered: B production, diffraction… strong efforts exist at both CDF and D Ø hard ME pdfs ME + PS hard radiation fragmentation to HF resummation soft radiation multiple parton scattering life at high luminosities…

3. Physics at LHC, 16 July 2004Marek Zieliński, Rochester3 The Fermilab Tevatron Collider lThe Tevatron is: è the highest-energy collider till LHC lIncreasing luminosity: èRun I (1992-95) ~0.1 fb -1 èRun IIa (2001~2005) ~1 fb -1 èRun IIb (2006-2009) ~4-8 fb -1 lStudies of QCD at highest Q 2 èPrecision tests of pQCD èPhenomenological models for “soft” aspects of QCD èTuning of Monte Carlo generators èProbing for new physics èUnderstanding backgrounds to many processes of interest

4. Physics at LHC, 16 July 2004Marek Zieliński, Rochester4 Tevatron Luminosity: Current and Future Approaching 10 32 cm -2 s -1 !  Tevatron has operated well in 2004  Already have >400 pb -1 of data on tape per experiment  Recent data taking rate ~10 pb -1 per week  Data taking efficiency 80-90%  Most results presented today are from first 130-210 pb -1  Much more to come by the time of LHC

5. Physics at LHC, 16 July 2004Marek Zieliński, Rochester5 Jet Physics at Tevatron lAt  s=1.96 TeV, cross section ~5x larger compared to Run I for jets with p T > 600 GeV èA jet factory… lHigher statistics important for: èbetter determination of proton structure at large x ètesting pQCD at a new level (resummation, NNLO theory, NLO event generators) ècontinued searches for new physics while testing distances ~10 -19 m  compositeness, W’, Z’, extra dimensions etc… lNew algorithms: èmidpoints èmassive jets, using jet p T Inclusive jet spectrum

6. Physics at LHC, 16 July 2004Marek Zieliński, Rochester6 Jet Definitions in Run II lRun I cone algorithm: èAdd up towers around a “seed” èIterate until stable èJet quantities: E T, ,  lModifications for Run II: èUse 4-vector scheme, p T instead of E T èAdd midpoints of jets as additional starting seeds èInfrared safe lCorrect to particles èUnderlying event, previous/extra interactions, energy loss out of cone due to showering in the calorimeter, detector response, resolution lCDF using the Run I JETCLU algorithm for some results, in the process of switching to midpoint lkT algorithm also used – see later Time “parton jet” “particle jet” “calorimeter jet” hadrons  CH FH EM

7. Physics at LHC, 16 July 2004Marek Zieliński, Rochester7 x-Q 2 Reach in Run I lDØ’s most complete jet cross section measurement in Run I ècovers |  | < 3.0 ècomplements HERA x-Q 2 range lUsed in CTEQ6 and MRST2001 fits to determine gluon at large x èEnhanced gluon at large x compared to previous fits CTEQ6M comparison: Big residual uncertainty for gluon PDF at high x 1<  <1.5 0.5<  <1 0<  <0.5 1.5<  <22<  <3

8. Physics at LHC, 16 July 2004Marek Zieliński, Rochester8 lFirst corrected Run II cross section for forward jets lImportant PDF information in cross section vs. rapidity lGood agreement between data and theory lLarge uncertainties due to jet energy scale èBig improvements already on the way Inclusive Jet Cross Section: Run II Midpoint

9. Physics at LHC, 16 July 2004Marek Zieliński, Rochester9 Central Inclusive Jet Cross Section: JETCLU lRapidity-dependent measurement in the works ● Run I reach extended by 150 GeV ● Data agree with NLO prediction within errors (Run I JETCLU used) ● Need to be corrected for hadronization/underlying event ● Watch the high p T -tail… Run II/Run I

10. Physics at LHC, 16 July 2004Marek Zieliński, Rochester10 Central Inclusive Jet Cross Section: kT lInclusive Jet Cross Section using kT algorithm èUses relative momentum of particles èNo split/merge ambiguities èInfrared and collinear safe lReasonable agreement between theory and data èNLO still needs to be corrected for hadronization and Underlying Event Cone jetk T jet D = Jet Size Parameter

11. Physics at LHC, 16 July 2004Marek Zieliński, Rochester11 kT-Jet Cross Section – Sensitive to UE? D = 1.0 Effect increases with increasing D Picking up Underlying Event? Not well modeled by NLO pQCD Pythia tuned to CDF Run I data (Pythia Tune A – see later) Good modeling of UE important

12. Physics at LHC, 16 July 2004Marek Zieliński, Rochester12 lCentral region |y jet | < 0.5, data sample ~143 pb -1 lRun II midpoint algorithm lAgrees within uncertainties with NLO/CTEQ6M lJet Energy Scale (<7%) -- dominant error on the measurement Dijet Production

13. Physics at LHC, 16 July 2004Marek Zieliński, Rochester13 b-Jet and bb Dijet Production lTest heavy-flavor production in QCD èProbe HF content of protons, g→bb splitting, flavor creation (back-to-back) è New Physics may show up as “bumps” in bb mass spectrum l 1 or 2 central tagged jets èFit mass distribution in the secondary vertex to b, c, uds templates lOngoing work, expect more to come Inclusive b-jet spectrum, 150 pb -1 Dijet Mass

14. Physics at LHC, 16 July 2004Marek Zieliński, Rochester14 Electro-Weak Bosons + Jets lA good testing ground for QCD  W/Z+n jets ~  s n in lowest order èPerturbation theory should be reliable  heavy boson ↔ large scale èNLO calculations available for up to 2 jets lW+jets, Z+jets èImportant backgrounds for other physics channels  Top, Higgs,…  +jet,  W/Z p T èTesting resummation techniques  Background to Higgs→  discovery channel at LHC lTesting ground for Monte Carlo tools required for precision measurements and searches for new physics èMulti-parton generators  Alpgen, MadGraph,… èNLO generators  MCFM, MC@NLO,... èCombining Parton-Shower and Matrix Element techniques to avoid “double counting”  MLM, CKKW,… prescriptions èTuning of ISR/FSR/MPI and soft Underlying Event important for comparisons to data lAll these aspects are being exercised/studied at the Tevatron, will benefit LHC physics

15. Physics at LHC, 16 July 2004Marek Zieliński, Rochester15 W + n Jets Cross Section vs n lTest of QCD predictions at large Q 2 ~M W 2 è fundamental channel for Top/Higgs/SUSY serches è Compared to LO Alpgen + Herwig + detector simulation lOne energetic and isolated electron + high E T jets lBackgrounds: Top dominates for 4-jet bin, QCD contributes to all jet bins Systematic uncertainty (10% in  1 to 40% in  4 ) limits the measurement sensitivity Results agree with LO QCD predictions within uncertainties The ratio R n/(n-1) measures the decrease in the cross section with the addition of one jet. It depends on  s

16. Physics at LHC, 16 July 2004Marek Zieliński, Rochester16 W + Jets Cross Section: Kinematics lReasonable agreement of E T and mass spectra with Alpgen + Herwig èSensitivity to variation of renormalization scale Highest E T jet in W + ≥1 jet Second highest E T jet in W + ≥2 jet, etc…

17. Physics at LHC, 16 July 2004Marek Zieliński, Rochester17 W + 2 Jets with b-tagging lData sample requires èa central electron with p T > 20 GeV è Missing E T > 25 GeV  2 jets: p T > 20 GeV, |  | < 2.5 èb-tagging based on impact parameter information lConsistent with Alpgen + Pythia èSeveral processes contribute èMass and ΔR distributions are sensitive to parton radiation process  ΔR is a measure of jet-jet distance in η-φ space lTowards the measurement of Wbb cross section and Higgs searches! MC: Alpgen+Pythia

18. Physics at LHC, 16 July 2004Marek Zieliński, Rochester18 Z + b Production Z+b signal observed at D Ø  Main background to search for associated HZ production Data 152  184 (ee) pb -1 :  p Tjet >20 GeV,  2.5  Secondary vertex tag Ratio (Z+b)/(Z+j)=0.024±0.007 consistent with NLO calculation lClean measurement of b-pdf at LHC? Useful for èSingle top: qb→qtW è Single top: gb→tW è (charged) Higgs+b: gb→Hb, H - t è Inclusive Higgs: bb→H 44% at Tevatron 83% at LHC Campbell et al. p Tjet > 15 GeV MC: Alpgen+Pythia

19. Physics at LHC, 16 July 2004Marek Zieliński, Rochester19 Diphoton Production lTesting NLO pQCD and resummation methods lSignature of interesting physics èOne of main Higgs discovery channels at LHC èPossible signature of GMSB SUSY Data: 2 isolated  s in central region, E T1,2 > 13, 14 GeV lGeneral agreement with NLO predictions, except èLow mass and high ΔΦ in DIPHOX (no resummation) èLow ΔΦ in RESBOS (resummation helps at large ΔΦ) èLO Pythia low by a factor ~2.2, but reasonable mass shape

20. Physics at LHC, 16 July 2004Marek Zieliński, Rochester20 Dijets: Azimuthal Decorrelations In 2→2 scattering, partons emerge back- to-back ➙ additional radiation introduces decorrelation in ΔΦ between the two leading partons/jets èSoft radiation: ΔΦ ~  èHard radiation: ΔΦ <  lΔΦ distribution is directly sensitive to higher-order QCD radiation lTesting fixed-order pQCD and parton- shower models across ΔΦ: èΔΦ~  :  Fixed-Order calculations unstable  Parton-Shower Monte Carlo’s applicable è2  /3 < ΔΦ <  :  First non-trivial description by 2→3 tree- level ME  2→3 NLO ME calculations became available recently (NLOJET++) èΔΦ < 2  /3 (3-jet “Mercedes”)  2→4 processes and higher Dijet production in lowest-order pQCD 3-jet production in lowest-order pQCD

21. Physics at LHC, 16 July 2004Marek Zieliński, Rochester21 ΔΦ: Comparison to Fixed-Order pQCD lΔΦ distribution: èSensitive to QCD radiation èNo need to reconstruct any other jets èReduced sensitivity to jet energy scale lData set ~150 pb -1 èCentral jets |y| < 0.5 èSecond-leading p T > 40 GeV lTowards larger p T, ΔΦ spectra more strongly peaked at ~  èIncreased correlation in ΔΦ lDistributions extend into the “4 final-state parton regime”, ΔΦ<2  /3 lLeading order (dashed blue curve) èDivergence at ΔΦ =  (need soft processes) èNo phase-space at ΔΦ<2  /3 (only three partons) lNext-to-leading order (red curve) èGood description by NLOJET++ over the whole range, except in extreme ΔΦ regions

22. Physics at LHC, 16 July 2004Marek Zieliński, Rochester22 ΔΦ: Comparison to Parton-Shower Monte Carlo’s lTesting the radiation process: è3 rd and 4 th jets generated by parton showers lHerwig 6.505 (default) èGood overall description! èSlightly too high in mid-range lPythia 6.223 (default) èVery different shape èToo steep dependence èUnderestimates low ΔΦ lΔΦ distributions are sensitive to the amount of initial-state radiation èPlot shows variation of PARP(67) from 1.0 (current default) to 4.0 (previous default, Tune A)  controls the scale of parton showers èIntermediate value suggested lMore Pythia tuning possible! CTEQ6L

23. Physics at LHC, 16 July 2004Marek Zieliński, Rochester23 Jet Shapes lJet shape: fractional energy flow  (r) = E T (r) / E T (R) lGoverned by multi-gluon emissions from the primary parton èTest of parton-shower model èSensitive to quark/gluon composition of final state èSensitive to underlying event lShapes are nearly identical for calorimeter towers and charged tracks Midpoint: R = 0.7 0.1 < |y| < 0.7 37 < p T < 380 GeV/c (1-Ψ)

24. Physics at LHC, 16 July 2004Marek Zieliński, Rochester24 Jet Shapes vs p T lp T fraction in outer part of cone (0.3 < R < 0.7) vs p T èJet shapes evolve from gluon to quark dominated profiles èData well described by Pythia Tune A and Herwig èDefault Pythia too narrow, especially at low p T

25. Physics at LHC, 16 July 2004Marek Zieliński, Rochester25 “Soft Aspects”: Underlying Event lUE contributes to hard-scatter processes èNot well understood theoretically èGood modeling essential l The studies: è Look at charged particle distributions (p T > 0.5 GeV, |  | < 1) relative to the leading jet (|  | < 2) è Focus on the region “Transverse” to the jet – high sensitivity to UE “hard” parton-parton collision:  outgoing jets with large p T  but: everything color-connected “Underlying Event”: everything but the two outgoing hard scattered “jets”  NOT the same as Min-Bias  Not independent of hard scatter (includes ISR/FSR/MPI)

26. Physics at LHC, 16 July 2004Marek Zieliński, Rochester26 UE: Data vs Monte Carlo lConsider particle density in the “Transverse” region èPoor desription by default Pythia èGood description by tuned Pythia (Tune A preferred by other studies) Default PythiaTuned Pythia

27. Physics at LHC, 16 July 2004Marek Zieliński, Rochester27 Emergence of Jets in Hard and MinBias Events lFind maximum p T particle in the “Transverse” region (rotated to 180º) èLabeled PTmaxT for Back-to-Back events, PTmax for MinBias èMeasure “associated” particle or PTsum density in ΔΦ relative to it as function of PTmaxT lObserve emergence of 3 rd jet as PTmaxT increases! èDensity shape the same in hard and MinBias events èPythia and Herwig: close but not exact Maximum p T particle in the “Transverse” region

28. Physics at LHC, 16 July 2004Marek Zieliński, Rochester28 ☛ √s dependence of the charged particle density for “Min-Bias” collisions compared with Pythia Tune A l Pythia Tune A predicts that 1% of all “Min-Bias” events at 1.8 TeV result from hard 2-to-2 parton-parton interactions with P T (hard) > 10 GeV/c è increases to 12% at 14 TeV l Work starting on “universal tuning” (Rick Field, CDF)  include jets, , Z, W, DY, HF etc… LHC? Using Tuned Pythia to Predict LHC ☛ Fraction of MinBias events with PT(hard) > 5 and 10 GeV vs √s, expected from PYTHIA Tune A

29. Physics at LHC, 16 July 2004Marek Zieliński, Rochester29 Summary Tevatron, CDF and D Ø are performing well èData samples already significantly exceed those of Run I èOn track for accumulating 4-8 fb -1 by 2009 lRobust QCD program is underway èJets, photons, W/Z+jets, heavy flavors  Jet energy scale is the dominant systematics – improvements on the way  Heavy flavor identification is working well  Probing hard scatter Matrix Elements to 10 -19 m,  s, pdfs, soft and hard radiation, jet structure, Underlying and MinBias Event properties èVerifying and tuning tools: NLO/NNLO calculations, Monte Carlo generators, resummation techniques, combining ME with PS  NLO does well for hard aspects  LO + Pythia give reasonable description of W/Z+n jets  Tuned Pythia models soft aspects well lQCD knowledge from Tevatron is essential for èPrecision measurements and searches for New Physics èExpectations for LHC

30. Physics at LHC, 16 July 2004Marek Zieliński, Rochester30 CDF and DØ Detectors  = 2  = 1 lNew silicon and drift chamber lUpgraded calorimeter and muon systems lUpgrade of Trigger/DAQ lNew silicon and fiber tracker lSolenoid (2 Tesla) lUpgrade of muon system lUpgrade of Trigger/DAQ

31. Physics at LHC, 16 July 2004Marek Zieliński, Rochester31 Run II High E T Jets E T = 666 GeV  = 0.43 E T = 633 GeV   = -0.19 Calorimeter LEGO Plot r-   view A high mass di-jet event: M jj = 1364 GeV /c 2 CDF

32. Physics at LHC, 16 July 2004Marek Zieliński, Rochester32  + b/c Cross Section lTest heavy-flavor production in QCD èProbe HF content of protons, g→bb splitting è Possible signatures of New Physics Data: 1 isolated  E T > 25 GeV, 1 jet with secondary vertex (“b/c-like”) èFit mass distribution in the secondary vertex to b, c, uds templates lQCD consistent with data èStill big uncertainties èNo new physics seen yet… cc bb

33. Physics at LHC, 16 July 2004Marek Zieliński, Rochester33 Central Inclusive Jet Cross Section: JETCLU lRapidity-dependent measurement in the works ● Run I reach extended by 150 GeV ● Data agree with NLO prediction within errors (Run I JETCLU used) Run II/Run I Data/Theory